Skip to main content

Full text of "Introduction to Free Software"

See other formats

inTrnoucTinn td 


j. EnnaaLEE aarai-inna 



j. mna HBrnanoaa 
d. mEGiRS JimanEE 

HCHDEm i i J 

Introduction to 
Free Software 

Jordi Mas Hernandez (coordinador) 
David Megias Jimenez (coordinador) 
Jesus M. Gonzalez-Barahona 
Joaquin Seoane Pascual 
Gregorio Robles 



GNUFDL* PID 00148366 

Introduction to Free Software 

Jordi Mas Hernandez 

Founding member of Softcatala and 
of the telematic network RedBBS. 
He has worked as a consultant in 
companies like Menta, Telepolis, 
Vodafone, Lotus, eresMas, Amena 
and Terra Espana. 

David Megias Jimenez 

Computer Science Engineer by the 
Universitat Autonoma de Barcelona 
(UAB, Spain). Master in Advanced 
Process Automatisation Techniques 
by the UAB. PhD. in Computer Sci- 
ence by the UAB. Associate Profes- 
sor in the Computer Science, Multi- 
media and Telecommunication De- 
partment of the Universitat Oberta 
de Catalunya (UOC, Spain) and Di- 
rector of the Master Programme in 
Free Software at the UOC. 

Jesus M. Gonzalez Barahona 

Associate professor in the Depart- 
ment of Telematic Systems and 
Computing of the Rey |uan Carlos 
University (Madrid, Spain), where 
he coordinates the GSyC/LibreSoft 
research group. His professional ar- 
eas of interest include the study of 
free software development and the 
transfer of knowledge in this field to 
the industrial sector. 

Joaquin Seoane Pascual 

Associate professor in the Depart- 
ment of Telematic Systems Engi- 
neering of the Technical Universi- 
ty of Madrid (Spain), where he has 
taught courses in programming, 
protocols, computer architecture, 
operating systems, Internet services, 
databases, systems administration 
and free software. His current inter- 
ests include the application of ICT in 
isolated regions of developing coun- 

Cregorio Robles 

Assistant professor at the Rey Juan 
Carlos University (Madrid, Spain), 
where he earned his PhD degree in 
February 2006. Besides his teaching 
tasks, he researches free software 
development from the point of view 
of software engineering, with spe- 
cial focus in quantitative issues. 

Third edition: September 2009 

Fundacio per a la Universitat Oberta de Catalunya. 

Av. Tibidabo, 39-43, 08035 Barcelona 

Material prepared by Eureca Media, SL 

© Jesus M. Gonzalez Barahona, Joaquin Seoane Pascual, Gregorio Robles 

Copyright © 2010, FUOC. Permission is granted to copy, distribute and modify this document either under the terms of the GNU 
Free Documentation Licence, Version 1 .2 or any subsequent version published by the Free Software Foundation, with no invariant 
sections or front-cover or back-cover texts, or under the terms of Creative Commons by-sa 3.0 license, at the option of the user. A 
copy of these licenses is included in the corresponding appendixes of this document. 


Software has become a strategic societal resource in the last few decades. 
The emergence of Free Software, which has entered in major sectors of 
the ICT market, is drastically changing the economics of software 
development and usage. Free Software - sometimes also referred to as 
"Open Source" or "Libre Software" - can be used, studied, copied, 
modified and distributed freely. It offers the freedom to learn and to 
teach without engaging in dependencies on any single technology 
provider. These freedoms are considered a fundamental precondition for 
sustainable development and an inclusive information society. 

Although there is a growing interest in free technologies (Free Software 
and Open Standards), still a limited number of people have sufficient 
knowledge and expertise in these fields. The FTA attempts to respond to 
this demand. 

Introduction to the FTA 

The Free Technology Academy (FTA) is a joint initiative from several 
educational institutes in various countries. It aims to contribute to a 
society that permits all users to study, participate and build upon existing 
knowledge without restrictions. 

What does the FTA offer? 

The Academy offers an online master level programme with course 
modules about Free Technologies. Learners can choose to enrol in an 
individual course or register for the whole programme. Tuition takes 
place online in the FTA virtual campus and is performed by teaching 
staff from the partner universities. Credits obtained in the FTA 
programme are recognised by these universities. 

Who is behind the FTA? 

The FTA was initiated in 2008 supported by the Life Long Learning 
Programme (LLP) of the European Commission, under the coordination 
of the Free Knowledge Institute and in partnership with three european 
universities: Open Universiteit Nederland (The Netherlands), Universitat 
Oberta de Catalunya (Spain) and University of Agder (Norway). 

For who is the FTA? 

The Free Technology Academy is specially oriented to IT professionals, 
educators, students and decision makers. 

What about the licensing? 

All learning materials used in and developed by the FTA are Open 
Educational Resources, published under copyleft free licenses that allow 
them to be freely used, modified and redistributed. Similarly, the 
software used in the FTA virtual campus is Free Software and is built 
upon an Open Standards framework. 

Evolution of this book 

The FTA has reused existing course materials from the Universitat 
Oberta de Catalunya and that had been developed together with 
LibreSoft staff from the Universidad Rey Juan Carlos. In 2008 this book 
was translated into English with the help of the SELF (Science, 
Education and Learning in Freedom) Project, supported by the 
European Commission's Sixth Framework Programme. In 2009, this 
material has been improved by the Free Technology Academy. 
Additionally the FTA has developed a study guide and learning activities 
which are available for learners enrolled in the FTA Campus. 


Users of FTA learning materials are encouraged to provide feedback and 
make suggestions for improvement. A specific space for this feedback is 
set up on the FTA website. These inputs will be taken into account for 
next versions. Moreover, the FTA welcomes anyone to use and distribute 
this material as well as to make new versions and translations. 

See for specific and updated information about the book, including 
translations and other formats: For 
more information and enrolment in the FTA online course programme, 
please visit the Academy's website: . 

I sincerely hope this course book helps you in your personal learning 
process and helps you to help others in theirs. I look forward to see you 
in the free knowledge and free technology movements! 

Happy learning! 

Wouter Tebbens 

President of the Free Knowledge Institute 
Director of the Free technology Academy 

© FUOC • XP07/M21 01 /02708 Introduction to Free Software 

The authors would like to thank the Foundation for the 
Universitat Oberta de Catalunya for financing the first edition 
of this work, and a large share of the improvements leading to 
the the second edition, as part of the Master Programme in 
Free Software offered by the University in question, where it is 
used as material for one of the subjects. 

The translation of this work into English has been made 
possible with the support from the SELF Project, the SELF 
Platform, the European Comission's programme on 
Information Society Technologies and the Universitat Oberta 
de Catalunya. We would like to thank the translation of the 
materials into English carried out by lexiaipark. 

The current version of these materials in English has been 
extended with the funding of the Free Technology Academy 
(FTA) project. The FTA project has been funded with support 
from the European Commission (reference no. 142706- 
LLP-l-2008-l-NL-ERASMUS-EVC of the Lifelong Learning 
Programme). This publication reflects the views only of the 
authors, and the Commission cannot be held responsible for 
any use which may be made of the information contained 

GNUR5L • PID 00148366 5 Introduction to Free Software 


"Anyone who hears this, if he can sing, may add and change at pleasure. Let it go from 
hand to hand: let those who request it have it. As a ball among young women, catch 
it if you can. 

Since this is of 'Good Love', lend it out gladly: do not make a mockery of its name by 
keeping it in reserve; nor exchange it for money by selling or renting it; for 'Good Love 1 
when bought, loses its charm." 

Juan Ruiz, Archpriest of Hita. The Book of Good Love (14* century, original in Ancient 

The first version of these notes was written by Jesus M. Gonzalez-Barahona, 
Joaquin Seoane Pascual and Gregorio Robles between April and September 
2003. Although we had been discussing for a while preparing a document of 
this type for the Free Software course that Joaquin and Jesus teach as part of 
the PhD programs of their respective departments, it was the initiative of the 
Open University of Catalonia (UOC) that commissioned us to prepare material 
to introduce the free software masters course, which finally encouraged us to 
get started. The involvement of Jordi Mas, the academic coordinator of the 
masters course, in this task was crucial, in that he proposed us for the job and 
put us in contact with the UOC, additionally supporting our relations with 
the UOC throughout the project's duration. 

Shortly after handing in the first edition, the authors started retouching the 
material as part of an ongoing process, although with varying degrees of ac- 
tivity, until this second edition was completed in May 2007. During this time, 
the first edition was extensively used in the free software masters of the UOC 
and in various other postgraduate courses in Spain and America. The expe- 
rience with the UOC has been followed with particular interest by Gregorio 
Robles, who has participated in it, and has therefore obtained feedback that 
has proven extremely valuable for improving the content. The three of us 
(Joaquin, Jesus, and since 2006, Gregorio) have also continued with the post- 
graduate software course at the UPM (Polytechnic University of Madrid) and 
at the URJC (Rey Juan Carlos University), taking advantage of it in order to 
test the material. 

Once again, the UOC has been the catalyser of this second edition, charging 
us with a commission that we have taken too long to complete. The work of 
Jordi Mas and David Megias (of the UOC) has been fundamental, and has pro- 
vided vital critical support for pushing forward this new edition. The work of 
Jose Ignacio Fernandez Villamor and Boni Garcia Gutierrez, pupils of Joaquin 
Seoane, who have collaborated in reviewing the materials for this second edi- 
tion, has also been essential. 

GNUFDL • PID_001 48366 6 Introduction to Free Software 

Previous materials 

Some of the texts in these notes are based on previous material, usually be- 
longing to the authors themselves, and in some cases to third parties (used 
with permission when not completely redrafted). Among them, we would like 
to mention the following (at the risk of omitting anyone important): 

• There are some fragments (especially on the chapters of history and the 
economy) inspired by the document "Free Software / Open Source: Infor- 
mation Society Opportunities for Europe?" [132], which Jesus Gonzalez- 
Barahona co-edited for the European Commission. However, the frag- 
ments in question have been extended, retouched and updated to such 
an extent that in many cases they may be difficult to recognise. 

• The section on monopolies and free software (section 5.4) has been based 
on the paper "Software libre, monopolios y otras yerbas" ("Free software, 
monopolies and other herbes") [84], by Jesus M. Gonzalez-Barahona. 

• The sections on legislative initiatives and public administration initiatives 
in relation to free software are partly based on "Iniciativas de las adminis- 
traciones publicas en relacion al Software Libre" ("Initiatives of public ad- 
ministrations related to free software") [103] (thanks to Pedro de las Heras 
for allowing us to use this material, which he co-authored). 

• Part of the section on motives for using free software in the public admin- 
istrations (Section 6.2) is based on the paper [85], by Jesus M. Gonzalez- 

• The chapter on free software engineering is an adaptation of the paper on 
the state of the art of software engineering applied to free software by Jesus 
M. Gonzalez-Barahona and Gregorio Robles for the magazines Novatica 
(Spanish version) and Upgrade (English version). 

• In the chapter on case studies, the part regarding the development of Lin- 
ux is based on a presentation made by Juan-Mariano de Goyeneche during 
the postgraduate course "Free Programs" of the Polytechnic University of 
Madrid during academic year 2002-03. 

• The historical part of the detailed study of GNOME has been taken from 
the historical introduction included in the book on "Applications devel- 
opment in GNOME2" prepared by GNOME Hispano and written by one 
of the authors of this book. 

• The FreeBSD case study is partly based on the presentation given by Jesus 
Rodriguez at the III HispaLinux Conference held in Madrid in the year 

GNUFDL • PID_001 48366 7 Introduction to Free Software 

• The Debian and Red Hat case studies are based on the previous work of 
Gonzalez-Barahona et al. who have reflected the results of the quantitative 
analysis of these two distributions in various papers. 

• Various materials, especially updates and new material in the chapter on 
case studies, were prepared by Jose Ignacio Fernandez Villamor and Boni 
Garcia Gutierrez towards the beginning of 2007 on a specific branch for 
modifications made in the context of that year's edition of the postgrad- 
uate subject of Joaquin Seoane at the UPM. A large proportion of those 
materials was included in time for the second edition. 

GNUR5L • PID 00148366 8 Introduction to Free Software 


Module 1 
Free Software 

Jesus M. Gonzalez-Barahona, Joaquin Seoane Pascual and Gregorio Robles 

1. Introduction 

2. A bit of history 

3. Legal aspects 

4. Developers and their motivations 

5. Economy 

6. Free software and public administrations 

7. Free software engineering 

8. Development environments and technologies 

9. Case studies 

10. Other free resources 

Module 2 

Jesus M. Gonzalez Barahona, Joaquin Seoane Pascual and Gregorio Robles 

1. Appendix A. Learning guide 

2. Appendix B. Key dates in the history of free software 

3. Appendix C. GNU Public License 

4. Appendix D. Texts of some legislative proposals and related documents 

5. Appendix E. Creative Commons' Attribution- ShareAlike 

6. Appendix F. GNU Free Documentation License 

GNUFDL • PID_001 48366 9 Introduction to Free Software 


ACM Association for Computing Machinery 

AFPL Aladdin Free Public License 

ALSA Advanced Linux Sound Architecture 

AOL America Online 

API Application program interface 

ARM Advanced RISC machines 

ASCII American standard code for information interchange 

AT&T American Telephone & Telegraph 

AITC Agency of Information Technologies and Communication 

ATK Accessibility Toolkit 

BIND Berkeley Internet Name Domain 

BIRT Business Intelligence and Reporting Tools 

BITNET Because It's There Network 

BSA Business Software Alliance 

BSD Berkeley Software Distribution 

BSDI Berkeley Software Design Incorporated 

BSI Bundesamt fur Sicherheit in der Informationstechnik 

CDDL Common Development and Distribution License 

CD-ROM Compact disc read-only memory 

CEPS Cisco Enterprise Print System 

CERN Conseil Europeen pour la Recherche Nucleaire 

CGI Common Gateway Interface 

COCOMO Cost construction model 

CORBA Common object request broker architecture 

CPL Common Public License 

CSRG Computer Systems Research Group 

CSS Cascading style sheet 

GNUFDL • PID_001 48366 1 Introduction to Free Software 

CVS Control version system 

DARPA Defense Advanced Research Projects Agency 

DBUS Desktop Bus 

DCOP Desktop communication protocol 

DEC Digital Equipment Corporation 

DECUS Digital Equipment Computer User Society 

DFSG Debian Free Software Guidelines 

DRM Digital rights management 

DSDP Device Software Development Platform 

DTD Document type definition 

DTP Data tools platform 

DVD Digital video disk 

ECTS European credit transfer scheme 

EMP Eclipse Modeling Project 

EPL Eclipse Public License 

HCEST Higher College of Experimental Sciences and Technology 

ETP Eclipse Tools Project 

FAQ Frequently asked questions 

FDL Free Documentation License 

FIC First International Computer 

FSF Free Software Foundation 

FTP File transfer protocol 

FUD Fear, uncertainty, doubt 

GCC GNU C Compiler 

GDB GNU Debugger 

GFDL GNU Free Documentation License 

GIMP GNU Image Manipulation Program 

GNAT GNU Ada Translator 

GNUFDL • PID_001 48366 1 1 Introduction to Free Software 

GNATS GNU Bug Tracking System 

GNU GNU's Not Unix 

GPL General Public License 

GTK GIMP Toolkit 

GUADEC GNOME User and Developer European Conference 

HIRD HURD of Interfaces Representing Depth 

HTML Hypertext markup language 

HTTP Hypertext transfer protocol 

HURD HIRD of Unix-Replacing Daemons 

R&D Research and development 

IBM International Business Machines Corporation 

IDE Integrated development environment 

IEC International Electrotechnical Commission 

IETF Internet Engineering Task Force 

INRIA Institut National de Recherche en Informatique et en Automatique 

IP Internet protocol 

IRC Internet Relay Chat 

ISO International Standards Organization 

ITU International Telecommunications Union 

JDK Java Developer Kit 

JPEG Joint Photographic Experts Group 

JRE Java Runtime Environment 

JVM Java Virtual Machine 

KBSt Koordinierungs-und Beratungsstelle der Bundesregierung fur Informa- 
tionstechnik in der Bundesverwaltung 

KDE K Desktop Environment 

LGPL Lesser General Public License 

LISP List processing language 

LLC Limited Liability Company 

GNUFDL • PID_001 48366 1 2 Introduction to Free Software 

IPA Intellectual Property Act 

LTS Long term support 

MCC Manchester City Council 

MIT Massachusetts Institute of Technology 

MPEG Moving Picture Experts Group 

MPL Mozilla Public License 

MTIC Mission Interministerielle de Soutin Technique pour le Developpe- 
ment des technologies de l'lnformation et de la Communication dans 

NASA National Aeronautics and Space Administration 

NCSA National Center for Supercomputing Applications 

NPL Netscape Public License 

NSFNet National Science Foundation Network 

NUMA Non-uniform memory access 

NYU New York University 

OASIS Organization for the Advancement of Structured Information Stan- 

ODF Open document format 

ODP Open Directory Project 

OHGPL OpenlPCore Hardware General Public License 

OLPC One Laptop Per Children 

WTO World Trade Organisation 

WIPO World Intellectual Property Organisation 

ORB Object request broker 

OSDN Open Software Development Network 

OSGi Open Services Gateway Initiative 

OSI Open Source Initiative 

GDD Gross Domestic Product 

PDA Portable digital assistant 

GNUFDL • PID_001 48366 1 3 Introduction to Free Software 

PDF Portable document format 

PDP Programmed data processor 

PHP PHP hypertext preprocessor 

PLOS Public Library of Science 

PNG Portable network graphics 

FAQ Frequently asked questions 

QPL Qt Public License 

RCP Rich client plaftorm 

RDF Resource description framework 

RFC Request for comments 

RFP Request for proposal 

RHAD Red Hat Advanced Development 

RPM Red Hat Package Manager 

RTF Rich text format 

SCO Santa Cruz Operation 

SPE Secretariat of Public Education 

SGI Silicon Graphics Incorporated 

SGML Standard generalised markup language 

SISSL Sun Industry Standards Source License 

SLS Softlanding Linux System 

SOA Service oriented architecture 

SPARC Scalable processor architecture 

SPICE Simulation program with integrated circuits emphasis 

SSL Secure socket layer 

TAMU Texas A&M University 

TCP Transport control protocol 

TEI Text Encoding Initiative 

TPTP Test and Performance Tools Project 

GNUFDL • PID_001 48366 1 4 Introduction to Free Software 

TRIPS Trade-related intellectual property rights 

UMTS Universal mobile telecommunications system 

UOC Open University of Catalonia 

USA United States of America 

USD United States dollar 

USENET User network 

USENIX Unix Users Group 

USL Unix System Laboratories 

UUCP UNIX to UNIX copy protocol 

VHDL Very high speed integrated circuit hardware description language 

W3C World Wide Web Consortium 

WIPO World Intellectual Property Organisation 

WTO World Trade Organisation 

WTP Web Tools Project 

WWW World Wide Web 

WYSIWYG What you see is what you get 

XCF Experimental computing facility format 

XML Extensible markup language 

Free Software 

Jesus M. Gonzalez-Barahona 
Joaquin Seoane Pascual 
Gregorio Robles 



Universitat Oberta 
de Catalunya 

GNURU- PID 00148386 Free Software 

Copyright © 2010, FUOC. Permission is granted to copy, distribute and modify this document either under the terms of the GNU 
Free Documentation Licence, Version 1 .2 or any subsequent version published by the Free Software Foundation, with no invariant 
sections or front-cover or back-cover texts, or under the terms of Creative Commons by-sa 3.0 license, at the option of the user. A 
copy of these licenses is included in the corresponding appendixes of this document. 

GNUFDL • PID 00148386 Free Software 


1. Introduction 9 

1.1. The concept of software freedom 9 

1.1.1. Definition 10 

1.1.2. Related terms 11 

1.2. Motivations 12 

1.3. The consequences of the freedom of software 12 

1.3.1. For the end user 13 

1.3.2. For the public administration 14 

1.3.3. For the developer 14 

1.3.4. For the integrator 15 

1.3.5. For service and maintenance providers 15 

1.4. Summary 15 

2. A bit of history 16 

2.1. Free software before free software 16 

2.1.1. And in the beginning it was free 17 

2.1.2. The 70s and early 80s 18 

2.1.3. The early development of Unix 19 

2.2. The beginning: BSD, GNU 20 

2.2.1. Richard Stallman, GNU, FSF: the free software 
movement is born 20 

2.2.2. Berkeley's CSRG 21 

2.2.3. The beginnings of the Internet 23 

2.2.4. Other projects 25 

2.3. Everything in its way 25 

2.3.1. The quest for a kernel 25 

2.3.2. The *BSD family 26 

2.3.3. GNU/Linux comes onstage 26 

2.4. A time of maturation 27 

2.4.1. End of the nineties 28 

2.4.2. Decade of 2000 31 

2.5. The future: an obstacle course? 38 

2.6. Summary 39 

3. Legal aspects 40 

3.1. Brief introduction to intellectual property 40 

3.1.1. Copyright 41 

3.1.2. Trade secret 43 

3.1.3. Patents and utility models 43 

3.1.4. Registered trademarks and logos 45 

3.2. Free software licences 45 

3.2.1. Types of licences 46 

GNUFDL' PID_001 48386 

3.2.2. Permissive licences 47 

3.2.3. Strong licences 50 

3.2.4. Distribution under several licences 54 

3.2.5. Program documentation 54 

3.3. Summary 56 

4. Developers and their motivations 57 

4.1. Introduction 57 

4.2. Who are developers? 57 

4.3. What do developers do? 58 

4.4. Geographical distribution 59 

4.5. Dedication 61 

4.6. Motivations 62 

4.7. Leadership 63 

4.8. Summary and conclusions 65 

5. Economy 66 

5.1. Funding free software projects 66 

5.1.1. Public funding 66 

5.1.2. Private not-for-profit funding 68 

5.1.3. Financing by someone requiring improvements 69 

5.1.4. Funding with related benefits 69 

5.1.5. Financing as an internal investment 70 

5.1.6. Other financing modes 71 

5.2. Business models based on free software 73 

5.2.1. Better knowledge 74 

5.2.2. Better knowledge with limitations 75 

5.2.3. Source of a free software product 76 

5.2.4. Product source with limitations 77 

5.2.5. Special licences 78 

5.2.6. Brand sale 79 

5.3. Other business model classifications 79 

5.3.1. Hecker classification 79 

5.4. Impact on monopoly situations 80 

5.4.1. Elements that favour dominant products 81 

5.4.2. The world of proprietary software 82 

5.4.3. The situation with free software 82 

5.4.4. Strategies for becoming a monopoly with free 

software 83 

6. Free software and public administrations 85 

6.1. Impact on the public administrations 85 

6.1.1. Advantages and positive implications 86 

6.1.2. Difficulties of adoption and other problems 89 

6.2. Actions of the public administrations in the world of free 

software 91 

Free Software 

GNUFDL • PID_001 48386 Free Software 

6.2.1. How to satisfy the needs of the public 
administrations? 91 

6.2.2. Promotion of the information society 93 

6.2.3. Research promotion 94 

6.3. Examples of legislative initiatives 95 

6.3.1. Draft laws in France 95 

6.3.2. Draft law of Brazil 96 

6.3.3. Draft laws in Peru 97 

6.3.4. Draft laws in Spain 98 

7. Free software engineering 100 

7.1. Introduction 100 

7.2. The cathedral and the bazaar 100 

7.3. Leadership and decision-making in the bazaar 102 

7.4. Free software processes 104 

7.5. Criticism of "The cathedral and the bazaar" 105 

7.6. Quantitative studies 106 

7.7. Future work 109 

7.8. Summary 110 

8. Development environments and technologies Ill 

8.1. Description of environments, tools and systems Ill 

8.2. Associated languages and tools 112 

8.3. Integrated development environments 113 

8.4. Basic collaboration mechanisms 113 

8.5. Source management 115 

8.5.1. CVS 116 

8.5.2. Other source management systems 119 

8.6. Documentation 120 

8.6.1. DocBook 122 

8.6.2. Wikis 122 

8.7. Bug management and other issues 123 

8.8. Support for other architectures 125 

8.9. Development support sites 126 

8.9.1. SourceForge 126 

8.9.2. SourceForge heirs 128 

8.9.3. Other sites and programs 128 

9. Case studies 129 

9.1. Linux 130 

9.1.1. A history of Linux 131 

9.1.2. Linux's way of working 132 

9.1.3. Linux's current status 133 

9.2. FreeBSD 135 

9.2.1. History of FreeBSD 135 

9.2.2. Development in FreeBSD 136 

9.2.3. Decision-making process in FreeBSD 136 

GNUFDL • PID_001 48386 Free Software 

9.2.4. Companies working around FreeBSD 137 

9.2.5. Current status of FreeBSD 138 

9.2.6. X-ray picture of FreeBSD 138 

9.2.7. Academic studies on FreeBDS 140 

9.3. KDE 140 

9.3.1. History of KDE 141 

9.3.2. Development of KDE 142 

9.3.3. The KDE League 142 

9.3.4. Current status of KDE 144 

9.3.5. X-ray picture of KDE 145 

9.4. GNOME 147 

9.4.1. History of GNOME 147 

9.4.2. The GNOME Foundation 148 

9.4.3. The industry working around GNOME 150 

9.4.4. GNOME's current status 151 

9.4.5. X-ray picture of GNOME 152 

9.4.6. Academic studies on GNOME 154 

9.5. Apache 154 

9.5.1. History of Apache 154 

9.5.2. Development of Apache 155 

9.5.3. X-ray picture of Apache 156 

9.6. Mozilla 157 

9.6.1. History of Mozilla 158 

9.6.2. X-ray picture of Mozilla 161 

9.7. 162 

9.7.1. 162 

9.7.2. Organisation of 163 

9.7.3. X-ray picture of 163 

9.8. Red Hat Linux 165 

9.8.1. History of Red Hat 165 

9.8.2. Current status of Red Hat 167 

9.8.3. X-ray of Red Hat 167 

9.9. Debian GNU/Linux 169 

9.9.1. X-ray picture of Debian 171 

9.9.2. Comparison with other operating systems 173 

9.10. Eclipse 174 

9.10.1. History of Eclipse 174 

9.10.2. Current state of Eclipse 175 

9.10.3. X-ray of Eclipse 176 

10. Other free resources 178 

10.1. The most important free resources 178 

10.1.1. Scientific papers 178 

10.1.2. Laws and standards 179 

10.1.3. Encyclopaedias 181 

10.1.4. Courses 182 

10.1.5. Collections and databases 183 

GNUFDL • PID_001 48386 Free Software 

10.1.6. Hardware 183 

10.1.7. Literature and art 184 

10.2. Licenses for other free resources 184 

10.2.1. GNU free documentation license 185 

10.2.2. Creative Commons licenses 186 

Bibliography 191 

GNUFDL» PID 00148386 

Free Software 

1. Introduction 

"If you have an apple and I have an apple and we exchange apples, then you and I will 
still each have one apple. But if you have an idea and I have an idea and we exchange 
these ideas, then each of us will have two ideas." 

Attributed to Bernard Shaw 

What is free software? What is it and what are the implications of a free pro- 
gram licence? How is free software developed? How are free software projects 
financed and what are the business models associated to them that we are ex- 
periencing? What motivates developers, especially volunteers, to become in- 
volved in free software projects? What are these developers like? How are their 
projects coordinated, and what is the software that they produce like? In short, 
what is the overall panorama of free software? These are the sort of questions 
that we will try to answer in this document. Because although free software is 
increasing its presence in the media and in debates between IT professionals, 
and although even citizens in general are starting to talk about it, it is still for 
the most part an unknown quantity. And even those who are familiar with it 
are often aware of just some of its features, and mostly ignorant about others. 

To begin with, in this chapter we will present the specific aspects of free soft- 
ware, focusing mainly on explaining its background for those approaching 
the subject for the first time, and underlining its importance. As part of this 
background, we will reflect on the definition of the term (to know what we 
are talking about) and on the main consequences of using (and the mere ex- 
istence of) free software. 

1.1. The concept of software freedom 

Since the early seventies we have become used to the fact that anyone com- 
mercialising a program can impose (and does impose) the conditions under 
which the program can be used. Lending to a third party may be prohibited 
for example. Despite the fact that software is the most flexible and adaptable 
item of technology that we have, it is possible to impose the prohibition (and 
it frequently is imposed) to adapt it to particular needs, or to correct its errors, 
without the explicit agreement of the manufacturer, who normally reserves 
the exclusive right to these possibilities. But this is just one of the possibilities 
that current legislation offers: free software, on the other hand, offers freedoms 
that proprietary software denies. 

Proprietary Software 

In this text we will use the 
term proprietary software to re- 
fer to any program that can- 
not be considered free soft- 
ware in accordance with the 
definition we provide later. 

GNUFDL» PID_001 48386 10 Free Software 

1.1.1. Definition 

So, the term free software, as conceived by Richard Stallman in his definition 
(Free Software Foundation, "Free software definition" 
philosophy/free-sw.html [120]), refers to the freedoms granted to its receiver, 
which are namely four: 

1) Freedom to run the program in any place, for any purpose and forever. 

2) Freedom to study how it works and to adapt it to our needs. This requires 
access to the source code. 

3) Freedom to redistribute copies, so that we can help our friends and neigh- 

4) Freedom to improve the program and to release improvements to the pub- 
lic. This also requires the source code. 

The mechanism that guarantees these freedoms, in accordance with current 
legislation, is distribution under a specific licence as we will see later on (chap- 
ter 3). Through the licence, the author gives permission for the receiver of the 
program to exercise these freedoms, adding also any restrictions that the au- 
thor may wish to apply (such as to credit the original authors in the case of a 
redistribution). In order for the licence to be considered free, these restrictions 
must not counteract the abovementioned freedoms. 

The ambiguity of the term free 

The English term free software includes the word free, standing for 'freedom', but the term 
can mean also 'free of charge' or 'gratis', which causes a great deal of confusion. Which is 
why in some cases the English borrow Spanish/French words and refer to libre software, 
as opposed to gratis software. 

Therefore, the definitions of free software make no reference to the fact that 
it may be obtained free of charge: free software and gratis software are two 
very different things. However, having said this, we should also explain that 
due to the third freedom, anyone can redistribute a program without asking 
for a financial reward or permission, which makes it practically impossible to 
obtain big profits just by distributing free software: anyone who has obtained 
free software may redistribute it in turn at a lower price, or even for free. 


Despite the fact that anyone can commercialise a given program at any price, and that 
this theoretically means that the redistribution price tends towards the marginal cost of 
copying the program, there are business models based precisely on selling free software, 
because there are many circumstances in which the consumer will be prepared to pay in 
exchange for certain other benefits, such as for example a guarantee, albeit a subjective 
one, for the software acquired or an added value in the choice, updating and organisation 
of a set of programs. 

GNUFDL • PID_001 48386 1 1 Free Software 

From a practical point of view, several texts define more precisely what con- 
ditions a licence must fulfil in order to be considered a free software licence. 
Among these, we would highlight for their historical importance, the free soft- 
ware definition of the Free Software Foundation ( 
ophy/free-sw.html) [120], the Debian guidelines for deciding whether a pro- 
gram is free (http://www.debian.Org/social_contract.html#guidelines) [104] 
and the definition of the term open source by the Open Source Initiative (http:/ 
/ [215], which is very simi- 
lar to the preceding ones. 


For example, the Debian guidelines go into the detail of allowing the author to demand 
that distributed source codes not be modified directly, but rather that the original is 
accompanied by separate patches and that binary programs be generated with different 
names to the original. They also demand that the licences do not contaminate other 
programs distributed by the same means. 

1.1.2. Related terms 

The term open source software, promoted by Eric Raymond and the Open 
Source Initiative is equivalent to the term free software . Philosophically speak- 
ing, the term is very different since it emphasises the availability of the 
source code and not its freedom, but the definition is practically the same 
as Debian's ("The open source definition", 1998 
docs/definition_plain.html) [183]. This name is politically more aseptic and 
emphasises the technical side, which can provide technical benefits, such as 
improved development and business models, better security, etc. Strongly crit- 
icised by Richard Stallman ("Why free software is better than open source") [204] 
and the Free Software Foundation ( [27], it has resonated 
far better with the commercial literature and with the company strategies that 
one way or another support the model. 

Other terms associated in some way to free software are as follows: 


These are gratis programs. They are normally only dis- 
tributed in binary format, and can be obtained free of 
charge. Sometimes it is possible to obtain permission to 
redistribute, and sometimes not, meaning that then it 
can only be obtained from the "official" site maintained 
for that purpose. It is frequently used to promote oth- 
er programs (normally with more complete functionali- 
ty) or services. Examples of this type of programs include 
Skype, Google Earth or Microsoft Messenger. 


This is not even gratis software, but rather a distribution 
method since usually the programs can be copied freely, 
generally without source code, but not used continuously 
without paying for them. The requirement to pay may be 
motivated by a limited functionality, being sent annoying 
messages or the mere appeal to the user's ethic. Also, the 
licence's legal terms may be used against the transgres- 

GNUFDL' PID 00148386 


Free Software 

Charityware, careware 

This is normally shareware that requires payment to be 
directed towards a sponsored charitable organisation. In 
many cases, instead of demanding payment, a voluntary 
contribution may be requested. Some free software, such 
as Vim, asks for voluntary contributions of this nature 
(Brian Molenaar, "What is the context of charityware?") 

Public domain 

Here, the author totally renounces all his rights in favour 
of the public domain, and this needs to be explicitly stat- 
ed in the program since otherwise, the program will be 
deemed proprietary and nothing can be done with it. In 
this case, if additionally the source code is provided, the 
program is free. 


This is a particular case of free software where the licence 
requires any distributed modifications to also be free. 

Proprietary, locked-in, non-free 

These are terms used to refer to software that is neither 
free nor open source. 

1.2. Motivations 

As we have seen, there are two large families of motivations for free software 
development, which likewise give rise to the two names by which it is known: 

• The ethical motivation, championed by the Free Software Foundation 
( [27], which has inherited the hacker culture and sup- 
ports the use of the term free, arguing that software is knowledge that 
should be shared unimpeded, that hiding it is antisocial and additionally 
claims that the ability to modify programs is a form of freedom of expres- 
sion. You can study this in more depth by reading (Free software, free society. 
Selected essays of Richard M. Stallmari) [211] or the analysis of Pekka Hima- 
nen (The hacker ethic and the spirit of the information age. Random House, 
2001) [144]. 

• The pragmatic motivation, championed by the Open Source Initiative 
( [54] which supports the use of the term open 
source, and argues the case of the technical and financial advantages that 
we will discuss in the next section. 

Aside from these two main motivations, people working on free software can 
do so for many other reasons, including for fun (Linus Torvalds and David 
Diamond, Texere, 2001) [217] or for money, potentially with sustainable busi- 
ness models. Chapter 4 studies these motivations in detail on the basis of ob- 
jective analyses. 

1.3. The consequences of the freedom of software 

Free software offers many advantages and, of the few disadvantages, many 
have been exaggerated (or invented) by proprietary competitors. The most 
well-founded disadvantage is the financial one, since as we have seen it is 
not possible to make much money from its distribution, which can and tends 

GNUFDL • PID_001 48386 1 3 Free Software 

to be made by someone other than the author. This is why other business 
models and financing mechanisms are needed, which we look into in chapter 
5. Other disadvantages, such as the lack of support or poor quality, are related 
to financing but also in many cases are false, since even software with no form 
of financing tends to offer good support levels thanks to user and developer 
forums, and often the quality is very high. 

Bearing in mind the financial considerations, we should note that the free 
software cost model is very different to the proprietary software cost model, 
since a large amount of it develops outside of the formal monetary economy, 
and frequently using exchange/barter mechanisms: "I give you a program that 
you are interested in, and you adapt it to your architecture and make the im- 
provements that you need." Chapter 7 discusses the right software engineering 
mechanisms to make the most of these unpaid for human resources with their 
own particular features, while chapter 8 studies the tools used to make this 
collaboration effective. Also, a large share of the costs is reduced by the fact 
that it is free, since new programs do not need to start from scratch, because 
they can reuse already made software. The distribution also has a much lower 
cost, since it is distributed via the Internet and with free advertising through 
public forums designed for this purpose. 

Another outcome of the freedoms is the quality resulting from the voluntary 
collaboration of people who contribute or discover and notify bugs in envi- 
ronments or situations that are unimaginable for the original developer. Plus, 
if a program does not offer sufficient quality, the competition may take it and 
improve on it on the basis of what there is. This is how collaboration and compe- 
tition, two powerful mechanisms, combine in order to produce better quality. 

Now let's examine the beneficial consequences for the receiver. 

1.3.1. For the end user 

The end user, whether an individual or a company, can find real competition 
in a market with a monopoly trend. To be precise, it does not necessarily de- 
pend on the software manufacturer's support, since there may be several com- 
panies, even small ones with the source code and the knowledge that allows 
them to do business while keeping certain programs free. 

Trying to find out the quality of a product no longer relies so much on the 
manufacturer's trustworthiness as on the guide given by the community's ac- 
ceptance and the availability of the source code. Also, we can forget about 
black boxes, that must be trusted "because we say so", and the strategies of 
manufacturers that can unilaterally decide whether to abandon or maintain 
a particular product. 

GNUFDL» PID_001 48386 14 Free Software 

Evaluating products before they are adopted has been made much easier now, 
since all we have to do is to install the alternative products in our real envi- 
ronment and test them, whereas for proprietary software we must rely on ex- 
ternal reports or negotiate tests with suppliers, which are not always possible. 

Because of the freedom to modify the program for own use, users are able 
to customise it or adapt it to own requirements correcting errors if there are 
any. The process of debugging errors found by proprietary software users is 
normally extremely laborious, if not impossible, since if we manage to get the 
errors debugged, the correction will often be incorporated in the following 
version, which may take years to be released, and which moreover we will have 
to buy again. With free software, on the other hand, we can make corrections 
or fixes ourselves, if we are qualified, or otherwise outsource the service. We 
can also, directly or by contracting external services, integrate the program 
with another one or audit its quality (for example in terms of security). To a 
great extent, control is passed on from the supplier to the user. 

1.3.2. For the public administration 

The public administration is a large user of special characteristics, as it has a 
special obligation towards its citizens, whether to provide accessible services, 
neutral in relation to manufacturers, or to guarantee the integrity, utility, pri- 
vacy and security of their data in the long term. All of the above makes it 
obligatory for the public administration to be more respectful towards stan- 
dards than private companies and to maintain data in open formats and to 
process data with software that is independent of usually foreign companies' 
strategies, certified as secure by an internal audit. Adaptation to standards is a 
notable feature of free software that proprietary software does not respect to 
the same extent, because it is generally eager to create captive markets. 

Also, the Administration serves as a sort of showcase and guide for industry, 
meaning that it has a great impact, which ought to be directed at weaving a 
technological fabric that generates national wealth. This wealth may be creat- 
ed by promoting the development of companies dedicated to developing new 
free software for the Administration, or maintaining, adapting or auditing ex- 
isting software. In chapter 6, we will look at this issue in more depth. 

1.3.3. For the developer 

For the software developer and producer, freedom significantly changes the 
rules of the game. It makes it easier to continue to compete while being small 
and to acquire cutting edge technology. It allows us to take advantage of oth- 
ers' work, competing even with another product by modifying its own code, 
although the copied competitor can then also take advantage of our code (if it 
is copyleft). If the project is well-managed, it is possible to obtain the free col- 
laboration of a large number of people and, also, to obtain access to a virtually 

GNUFDL» PID_001 48386 15 Free Software 

free and global distribution system. Nonetheless, the issue of how to obtain 
financial resources remains, if the software is not the product of a paid-for 
commission. Chapter 5 deals with this aspect. 

1.3.4. For the integrator 

For integrators, free software is paradise. It means that there are no longer 
black boxes that need to be fitted together, often using reverse engineering. 
Rough edges can be smoothed out and parts of programs can be integrated in 
order to obtain the required integrated product, because there is a huge shared 
pool of free software from which the parts can be extracted. 

1.3.5. For service and maintenance providers 

Having the source code changes everything and puts us in the same position 
as the producer. If the position is not the same, it is because we are lacking an 
in-depth knowledge of the program that only the developer has, which means 
that it is advisable for maintenance providers to participate in the projects 
that they are required to maintain. The added value of services is much more 
appreciated because the cost of the program is low. It is currently the clearest 
business with free software and the one where the most competition is pos- 

1.4. Summary 

This first chapter has served as a preliminary encounter with the world of free 
software. The concept defined by Richard Stallman is based on four freedoms 
(freedom to execute, freedom to study, freedom to redistribute and freedom 
to improve), two of which require access to the source code. This accessibility 
and its advantages have motivated another less ethical and more pragmatic 
point of view, defended by the Open Source Initiative, which has given rise to 
another term: open source software . We have also mentioned other related sim- 
ilar or opposite terms, which serve to clarify various concepts. Finally, we have 
discussed the consequences of free software for the main parties involved. 

GNUFDL" PID 00148386 16 Free Software 

2. A bit of history 

"When I started working at the MIT Artificial Intelligence Lab in 1971, I became part 
of a software-sharing community that had existed for many years. Sharing of software 
was not limited to our particular community; it is as old as computers, just as sharing of 
recipes is as old as cooking. But we did it more than most. [...] We did not call our software 
free software, because that term did not yet exist; but that is what it was. Whenever people 
from another university or a company wanted to port and use a program, we gladly 
let them. If you saw someone using an unfamiliar and interesting program, you could 
always ask to see the source code, so that you could read it, change it, or cannibalize 
parts of it to make a new program." 

Richard Stallman, "The GNU Project" (originally published in the book Open sources) [208] 

Although all the histories associated to IT are necessarily brief, free software's 
is one of the longest. In fact, we could say that in the beginning almost all 
developed software fulfilled the definition of free software, even though the 
concept didn't even exist yet. Later the situation changed completely, and 
proprietary software dominated the scene, almost exclusively, for a fairly long 
time. It was during that period that the foundations were laid for free software 
as we know it today, and when bit by bit free programs started to appear. Over 
time, these beginnings grew into a trend that has progressed and matured 
to the present day, when free software is a possibility worth considering in 
virtually all spheres. 

This history is largely unknown, to such an extent that for many IT profes- 
sionals proprietary software is software "in its natural state". However, the sit- 
uation is rather the opposite and the seeds of change that could first be dis- 
cerned in the first decade of the 21 st century had already been sown in the 
early 1980s. 


There are not many detailed histories of free software, and the ones that there are, are 
usually papers limited to their main subject. In any case, interested readers can extend 
their knowledge of what we have described in this chapter by reading "Open Source Ini- 
tiative. History of the OSI" [146] (, which 
emphasises the impact of free software on the business community in the years 1998 
and 1999; "A brief history of free/open source software movement" [190], by Chris Rasch, 
which covers the history of free software up until the year 2000, or "The origins and fu- 
ture of open source software" (1999) [177], by Nathan Newman, which focuses to a large 
extent on the US Government's indirect promotion of free software or similar systems 
during the decades of the 1970s and the 1980s. 

2.1. Free software before free software 

Free software as a concept did not appear until the beginning of the 1980s. 
However, its history can be traced back to several years earlier. 

GNUFDL • PID_001 48386 1 7 Free Software 

2.1.1. And in the beginning it was free 

During the sixties, the IT panorama was dominated by large computers, main- 
ly installed in companies and governmental institutions. IBM was the leading 
manufacturer, way ahead of its competition. During this period, when buying 
a computer (the hardware), the software came added. As long as the mainte- 
nance contract was paid for, access was given to the manufacturer's software 
catalogue. Plus, the idea of programs being something "separate" from a com- 
mercial point of view was uncommon. 

In this period, software was normally distributed together with its source code 
(in many cases just as source code), and in general, with no practical restric- 
tions. User groups such as SHARE (users of IBM systems) or DECUS (DEC users) 
participated in these exchanges, and to a certain extent, organised them. The 
"Algorithms" section of the magazine Communications of the ACM was another 
good example of an exchange forum. We could say that during these early 
years of IT, software was free, at least in the sense that those who had access 
to it could normally have access to the source code, and were used to sharing 
it, modifying it and also sharing these modifications. 

On 30 th June 1969, IBM announced that as of 1970, it would sell part of its soft- 
ware separately (Burton Grad, 2002) [131]. This meant that its clients could no 
longer obtain the programs they needed included in the price of the hardware. 
Software started to be perceived as something with an intrinsic value, and 
consequently, it became more and more common to scrupulously restrict ac- 
cess to the programs and the possibility of users sharing, modifying or study- 
ing the software was limited as much as possible (technically and legally). In 
other words, the situation changed to the one that continues to be case in the 

world of software at the beginning of the 21 st century. 


Readers interested in learning about this transition period, can read, for example "How 
the ICP Directory began" [226] (1998), in which Larry Welke discusses how one of the 
first software catalogues not associated to a manufacturer was born, and how during this 
process it was discovered that companies would be prepared to pay for programs not 
made by their computer manufacturers. 

In the mid-1970s it was already totally common, in the field of IT, to find 
proprietary software. This meant an enormous cultural change among profes- 
sionals who worked with software and was the beginning of a flourishing of 
a large number of companies dedicated to this new business. It would still be 
almost a decade before what we now know as free software started to appear in 
an organised manner and as a reaction to this situation. 

GNUFDL" PID_001 48386 18 Free Software 

2.1.2. The 70s and early 80s 

Even when the overwhelming trend was to explore the proprietary software 
model, there were initiatives that showed some of the characteristics of what 
would later be considered free software. In fact, some of them produced free 
software as we would define it today. Of these, we would mention SPICE, TeX 
and Unix, which is a much more complex case. 

SPICE (Simulation Program with Integrated Circuit Emphasis) is a program 
developed by the University of California, in Berkeley, in order to simulate the 
electrical characteristics of an integrated circuit. It was developed and placed 
in the public domain by its author, Donald O. Pederson, en 1973. SPICE was 
originally a teaching tool, and as such rapidly spread to universities world- 
wide. There it was used by many students of what was then an emerging dis- 
cipline: integrated circuits design. Because it was in the public domain, SPICE 
could be redistributed, modified, studied. It could even be adapted to specific 
requirements, and that version could be sold as a proprietary product (which 
is what a large number of companies have done dozens of times throughout 
their history). With these characteristics, SPICE had all the cards to become 
the industry standard, with its different versions. And indeed, that is what 
happened. This was probably the first program with free software characteris- 
tics that for a certain period captured a market, the one of integrated circuits 
simulators, and that undoubtedly was able to do so precisely thanks to these 
characteristics (in addition to its undeniable technical qualities). 


More information on the history of SPICE can be consulted in "The life of SPICE", pre- 
sented during the Bipolar Circuits and Technology Meeting, Minneapolis, MN, USA, in 
September 1996 [175]. 

You can find the SPICE web page at 

Donald Knuth started to develop TeX during a sabbatical year, in 1978. TeX is 
an electronic typography system commonly used for producing high-quality 
documents. From the start, Knuth used a licence that today would be consid- 
ered a free software licence. When the system was considered sufficiently sta- 
ble, in 1985, he maintained that licence. At that time, TeX was on the largest 
and most well-known systems that could be considered free software. 


You can find some of the milestones in the history of TeX by consulting online http:// [39]. For further details, the cor- 
responding article in Wikipedia is also extremely useful, 
TeX [233]. 

GNUFDL" PID_001 48386 19 Free Software 

2.1.3. The early development of Unix 

Unix, one of the first portable operating systems, was originally created by 
Thompson and Ritchie (among others) from AT&T's Bell Labs. It has contin- 
ued to develop since its birth around 1972, giving rise to endless variants sold 
(literally) by tens of companies. 

In the years 1973 and 1974, Unix arrived at many universities and research 
centres worldwide, with a licence that permitted its use for academic purpos- 
es. Although there were certain restrictions that prevented its free distribu- 
tion, among the organisations that did possess a licence the functioning was 
very similar to what would later be seen in many free software communities. 
Those who had access to the Unix source code were dealing with a system 
that they could study, improve on and extend. A community of developers 
emerged around it, which soon gravitated towards the CSRG of the University 
of California, in Berkeley. This community developed its own culture, which 
as we will see later, was very important in the history of free software. Unix 
was, to a certain extent, an early trial for what we would see with GNU and 
Linux several years later. It was confined to a much smaller community, and 
the AT&T licence was necessary, but in all other aspects, its development was 
very similar (in a far less communicated world). 

Development methods inherent to free software 

In Netizens. On the history and impact of Usenet and the Internet (IEEE Computer Society 
Press, 1997 [139], page 139) we can read a few lines that could refer to many free soft- 
ware projects: "Contributing to the value of Unix during its early development, was the 
fact that the source code was open and available. It could be examined, improved and 

Page 142 of the same work states the following: "Pioneers like Henry Spencer agree on 
how important it was to those in the Unix community to have the source code. He 
notes how having the sources made it possible to identify and fix the bugs that they 
discovered. [...] Even in the late 1970s and early 1980s, practically every Unix site had 
complete sources". 

The text of Marc Rochkind "Interview with Dick Haight" is even more explicit (Unix 
Review, May 1986) [198]: "that was one of the great things about Unix in the early days: 
people actually shared each other's stuff. [...] Not only did we learn a lot in the old days 
from sharing material, but we also never had to worry about how things really worked 
because we always could go read the source." 

Over time, Unix also became an early example of the problems that could arise 
from proprietary systems that at first sight "had some free software feature". 
Towards the end of the 1970s and especially during the decade of the 1980s, 
AT&T changed its policy and access to new versions of Unix became difficult 
and expensive. The philosophy of the early years that had made Unix so pop- 
ular among developers, changed radically to such an extent that in 1991 AT&T 
even tried to sue the University of Berkeley for publishing the Unix BSD code 
that Berkeley's CSRG had created. But this is another story that we will pick 
up on later. 

GNUFDL» PID_001 48386 20 Free Software 

2.2. The beginning: BSD, GNU 

All of the cases discussed in the previous section were either individual initia- 
tives or did not strictly comply with the requirements of free software. It was 
not until the beginning of the 1980s that the first organised and conscious 
projects to create systems comprising free software appeared. During that pe- 
riod, the ethical, legal and even financial grounds of these projects started to 
be established (probably more importantly), with them being developed and 
completed right up to the present day. And since the new phenomenon need- 
ed a name, this was when the term free software was first minted. 

2.2.1. Richard Stallman, GNU, FSF: the free software movement 
is born 

At the beginning of 1984, Richard Stallman, who at the time was employed 
by the MIT AI Lab, quited his job to started working on the GNU project. 
Stallman considered himself to be a hacker of the kind that enjoys sharing his 
technological interests and his code. He didn't like the way that his refusal to 
sign exclusivity or non-sharing agreements made him an outcast in his own 
world, and how the use of proprietary software in his environment left him 
impotent in the face of situations that could easily be resolved before. 

His idea when he left the MIT was to build a complete software system, for gen- 
eral use, but totally free ("The GNU Project", DiBona et al.) [208]. The system 
(and the project that would be responsible for making it come true) was called 
GNU ("GNU's not Unix", a recursive acronym). Although from the beginning 
the GNU project included software in its system that was already available (like 
TeX or, later, the X Window system), there was still a lot to be built. Richard 
Stallman started by writing a C compiler (GCC) and an editor (Emacs), both 
of which are still in use today (and very popular). 

From the start of the GNU project, Richard Stallman was concerned about the 
freedoms that the users of the software would have. He wanted not only those 
who received programs directly from the GNU project to continue to enjoy 
the same rights (modification, redistribution, etc.) but also those who received 
it after any number of redistributions and (potentially) modifications. For this 
reason he drafted the GPL licence, probably the first software licence designed 
specifically in order to guarantee that a program would be free in this way. 
Richard Stallman called the generic mechanism that these GPL type licences 
use in order to achieve these guarantees, copyleft, which continues to be the 
name of a large family of free software licences (Free Software Foundation, 
GNU General Public Licence, version 2, June 1991) [118]. 

GNUFDL» PID_001 48386 21 Free Software 

Richard Stallman also founded the Free Software Foundation (FSF) in order 
to obtain funds, which he uses to develop and protect free software, and es- 
tablished his ethical principles with the "The GNU Manifesto" (Free Software 
Foundation, 1985) [117] and "Why software should not have owners" (Richard 
Stallman, 1998) [207]. 

From a technical point of view, the GNU project was conceived as a highly 
structured endeavor with very clear goals. The usual methodology was based 
on relatively small groups of people (usually volunteers) developing one of the 
tools that would then fit perfectly into the complete jigsaw (the GNU system). 
The modularity of Unix, on which this project was inspired, fully coincided 
with that idea. The working method generally implied the use of Internet, but 
because at that time it was not extensively implanted, the Free Software Foun- 
dation would also sell tapes on which it would record the applications, which 
means that it was probably one of the first organisations to obtain financial 
compensation (albeit in a rather limited way) from creating free software. 

In the early 90s, about six years after the project was founded, GNU was very 
close to having a complete system similar to Unix. However, at that point it 
had not yet produced one of the key parts: the system's core (also known as the 
kernel, the part of the operating system that handles with the hardware, ab- 
stracts it, and allows applications to share resources, and essentially, to work). 
However, GNU software was very popular among the users of several different 
variants of Unix, at the time the most commonly used operating system in 
businesses. Additionally, the GNU project had managed to become relatively 
well known among IT professionals, and especially among those working at 
universities. In that period, its products already had a well-deserved reputa- 
tion for stability and good quality. 

2.2.2. Berkeley's CSRG 

Since 1973, the CSRG (Computer Science Research Group) of the University 
of California at Berkeley had been one of the centres where most of the Unix- 
related developments had been made, especially between 1979 and 1980. Not 
only were applications ported and other new ones built to run on Unix, but 
also important improvements were made to the kernel and a lot of function- 
ality had been added. For example, during the 80s, several DARPA contracts 
(under the US Department of Defence) financed the implementation of TCP/ 
IP which until today has been considered the reference for the protocols that 
make the Internet work (in the process, linking the development of the In- 
ternet and the expansion of Unix workstations). Many companies used the 
CSRG's developments as the bases for their Unix versions giving rise to well- 
known systems at the time, such as SunOS (Sun Microsystems) or Ultrix (Dig- 
ital Equipment). This is how Berkeley became one of the two fundamental 
sources of Unix, together with the "official" one, AT&T. 

GNUFDL" PID_001 48386 22 Free Software 

In order to use all of the code that the CSRG produced (and the code of the 
collaborators of the Unix community which to some extent they coordinated), 
it was necessary to have AT&T's Unix licence, which was becoming increas- 
ingly difficult (and expensive) to obtain, especially if access to the system's 
source code was required. Partly in an attempt to overcome this problem, in 
June 1989 the CSRG released the part of Unix associated to TCP/IP (the im- 
plementation of the protocols in the kernel and the utilities), which did not 
include AT&T code. It was called the Networking Release 1 (Net-1). The licence 
with which it was released was the famous BSD licence, which except for 
certain problems with its clauses on advertising obligations, has always been 
considered an example of a minimalist free software licence (which in addi- 
tion to allowing free redistribution, also allows incorporation into proprietary 
products). In addition, the CSRG tested a novel financing model (which the 
FSF was already trying out successfully): it sold tapes with its distribution for 
USD 1,000 each. Despite the fact that anybody in turn could redistribute the 
content of the tapes without any problem (because the licence allowed it), 
the CSRG sold tapes to thousands of organisations thus obtaining funds with 
which to continue developing. 

Having witnessed the success of the Net-1 distribution, Keith Bostic proposed 
to rewrite all of the code that still remained from the original AT&T Unix. 
Despite the scepticism of some members of the CSRG, he made a public an- 
nouncement asking for help to accomplish this task, and little by little the util- 
ities (rewritten on the basis of specifications) became integrated into Berkeley's 
system. Meanwhile, the same process was done with the kernel, in such a way 
that most of the code that had not been produced by Berkeley or volunteer 
collaborators was rewritten independently. In June 1991, after obtaining per- 
mission from the University of Berkeley's governing body Networking Release 2 
(Net- 2) was distributed, with almost all of the kernel's code and all of the util- 
ities of a complete Unix system. The set was once again distributed under the 
BSD licence and thousands of tapes were sold at a cost of USD 1,000 per unit. 

Just six months after the release of Net-2, Bill Jolitz wrote the code that 
was missing for the kernel to function on the i386 architecture, releasing 
386BSD, which was distributed over the Internet. On the basis of that code 
later emerged, in succession, all the systems of the *BSD family: first NetBSD 
appeared, as a compilation of the patches that had been contributed over the 
Net in order to improve 386BSD; later FreeBSD appeared, as an attempt to 
focus on the support of the i386 architecture; several years later the OpenB- 
SD project was formed, with an emphasis on security. And there was also a 
proprietary version based on Net-2 (although it was certainly original, since it 
offered its clients all the source code as part of the basic distribution), which 
was done independently by the now extinct company BSDI (Berkeley Software 
Design Inc.). 

GNUFDL» PID_001 48386 23 Free Software 

Partly as a reaction to the distribution produced by BSDI, the AT&T subsidiary 
that held the Unix licence rights, Unix System Laboratories (USL), tried to sue 
first BSDI and then the University of California. The accusation was that the 
company had distributed its intellectual property without permission. Follow- 
ing various legal manoeuvres (which included a countersuit by the University 
of California against USL), Novell bought the Unix rights from USL, and in 
January 1994 reached an out-of-court settlement with the University of Cali- 
fornia. As a result of this settlement, the CSRG distributed version 4.4BSD-Lite, 
which was soon used by all the projects of the *BSD family. Shortly afterwards 
(after releasing version 4.4BSD-Lite Release 2), the CSRG disappeared. At that 
point, some feared that it would be the end of *BSD systems, but time has 
shown that they are still alive and kicking under a new form of management 
that is more typical of free software projects. Even in the first decade of the 
year 2000 the projects managed by the *BSD family are among the oldest and 
most consolidated in the world of free software. 


The history of Unix BSD is illustrative of a peculiar way of developing software during 
the seventies and eighties. Whoever is interested in it can enjoy reading "Twenty years of 
Berkeley Unix" (Marshall Kirk McKusick, 1999) [170], which follows the evolution from 
the tape that Bob Fabry took to Berkeley with the idea of making one of the first versions 
of Thompson and Ritchie's code function on a PDP-1 1 (bought jointly by the faculties of 
informatics, statistics and mathematics), through to the lawsuits filed by AT&T and the 
latest releases of code that gave rise to the *BSD family of free operating systems. 

2.2.3. The beginnings of the Internet 

Almost since its creation in the decade of the 1970s, Internet has been closely 
related to free software. On the one hand, since the beginning, the communi- 
ty of developers that built the Internet had several clear principles that would 
later become classics in the world of free software; for example, the impor- 
tance of users being able to help fix bugs or share code. The importance of 
BSD Unix in its development (by providing during the eighties the most pop- 
ular implementation of the TCP/IP protocols) made it easy to transfer many 
habits and ways of doing things from one community - the developers cen- 
tred around the CSRG - to another community - the developers who were 
building what at the time was NSFNet and would later become Internet - and 
vice versa. Many of the basic applications for the Internet's development, such 
as Sendmail (mail server) or BIND (implementation of the domain name ser- 
vices) were free and, to a great extent, the outcome of collaboration between 
these two communities. 

Finally, towards the end of the 80s and in the decade of the 90s, the free soft- 
ware community was one of the first to explore in depth the new possibilities 
offered by the Internet for geographically disperse groups to collaborate. To a 
large extent, this exploration made the mere existence of the BSD community 
possible, the FSF or the development of GNU/Linux. 

GNUFDL' PID 00148386 


Free Software 

One of the most interesting aspects of the Internet's development, from the 
free software point of view, was the completely open management of its docu- 
ments and its rules. Although it may seem normal today (because it is custom- 
ary, for example, in the IETF or the World Wide Web Consortium), at the time, 
the free availability of all its specifications, and design documents including 
the norms that define the protocols, was something revolutionary and funda- 
mental for its development. In Netizens. On the history and impact of Usenet and 
the Internet [139] (page 106) we can read: 

"This open process encouraged and led to the exchange of information. Technical devel- 
opment is only successful when information is allowed to flow freely and easily between 
the parties involved. Encouraging participation is the main principle that made the de- 
velopment of the Net possible." 

We can see why this paragraph would almost certainly be supported by any 
developer referring to the free software project in which he is involved. 

In another quote, on "The evolution of packet switching" [195] (page 267) we 
can read: 

"Since ARPANET was a public project connecting many major universities and research 
institutions, the implementation and performance details were widely published." 

Obviously, this is what tends to happen with free software projects, where all 
the information related to a project (and not only to its implementation) is 
normally public. 

In this atmosphere, and before the Internet, well into the nineties, became 
an entire business, the community of users and its relationship with develop- 
ers was crucial. During that period many organisations learned to trust not 
a single supplier of data communication services, but rather a complex com- 
bination of service companies, equipment manufacturers, professional devel- 
opers, and volunteers, etc. The best implementations of many programs were 
not those that came with the operating system purchased together with the 
hardware, but rather free implementations that would quickly replace them. 
The most innovative developments were not the outcome of large company 
research plans but rather the product of students or professionals who tested 
ideas and collected feedback sent to them by various users of their free pro- 

As we have already mentioned, Internet also offered free software the basic 
tools for long-distance collaboration. Electronic mail, news groups, anony- 
mous FTP services (which were the first massive stores of free software) and, 
later, the web-based integrated development systems have been fundamental 
(and indispensable) for the development of the free software community as 
we know it today, and in particular, for the functioning of the immense ma- 
jority of free software projects. From the outset, projects such as GNU or BSD 


Readers interested in the 
evolution of the Internet, 
written by several of its key 
protagonists, can consult 
"A brief history of the Inter- 
net" (published by the ACM, 
1997) [166]. 

GNUFDL» PID_001 48386 25 Free Software 

made massive and intensive use of all these mechanisms, developing, at the 
same time as they used them, new tools and systems that in turn improved 
the Internet. 

2.2.4. Other projects 

During the 1980s many other important free software projects saw the light 
of day. We highlight for their importance and future relevance, X Window 
(windowing system for Unix-type systems), developed at the MIT, one of the 
first examples of large-scale funding for a free project financed by a business 
consortium. It is also worth mentioning Ghostscript, a PostScript document 
management system developed by a company called Aladdin Software, which 
was one of the first cases of searching for a business model based on producing 
free software. 

Towards the end of the 1980s, there was already an entire constellation of 
small (and not so small) free software projects underway. All of them, together 
with the large projects we have mentioned up until now, established the bases 
of the first complete free systems, which appeared in the beginning of the 

2.3. Everything in its way 

Around 1990, most of the components of a complete system were ready as free 
software. On the one hand, the GNU project and the BSD distributions had 
completed most of the applications that make up an operating system. On 
the other hand, projects such as X Window or GNU itself had built from win- 
dowing environments to compilers, which were often among the best in their 
class (for example, many administrators of SunOS or Ultrix systems would re- 
place their system's proprietary applications for the free versions of GNU or 
BSD for their users). In order to have a complete system built exclusively with 
free software, just one component was missing: the kernel. Two separate and 
independent efforts came to fill the gap: 386BSD and Linux. 

2.3.1. The quest for a kernel 

Towards the end of the 1980s and beginning of the 1990s, the GNU project 
had a basic range of utilities and tools that made it possible to have a complete 
operating system. Even at the time, many free applications, including the par- 
ticularly interesting case of X Window, were the best in their field (Unix util- 
ities, compilers...). However, to complete the jigsaw a vital piece was missing: 
the operating system's kernel. The GNU project was looking for that missing 
piece with a project known as Hurd, which intended to build a kernel using 
modern technologies. 

GNUFDL» PID_001 48386 26 Free Software 

2.3.2. The *BSD family 

Practically at the same time, the BSD community was also on the path towards 
a free kernel. The Net- 2 distribution was only missing six files in order to com- 
plete it (the rest had already been built by the CSRG or its collaborators). In 
the beginning of 1992, Bill Jolitz finished those files and distributed 386BSD, 
a system that functioned on the i386 architecture and that in time would give 
rise to the projects NetBSD, FreeBSD and OpenBSD. Progress in the following 
months was fast, and by the end of the year it was sufficiently stable to be 
used in non-critical production environments, which included, for example, 
a windows environment thanks to the XFree project (which had provided X 
Window for the i386 architecture) or a great quality compiler, GCC. Although 
there were components that used other licences (such as those from the GNU 
projects, which used the GPL), most of the system was distributed under the 
BSD licence. 


Some episodes of this period illustrate the capability of the free software development 
models. There is the well-known case of Linus Torvalds, who developed Linux while a 
second-year student at the University of Helsinki. But this is not the only case of a student 
who made his way thanks to his free developments. For example, the German Thomas 
Roel ported X11R4 (a version of the X Window system) to a PC based on a 386. This 
development took him to work at Dell, and later to become the founder of the X386 and 
XFree projects, which were fundamental for quickly giving GNU/Linux and the *BSDs a 
windows environment. You can read more about the history of XFree and Roel's role in 
it in "The history of xFree86" (Linux Magazine, December 1991) [135]. 

Then came the lawsuit from USL, which made many potential users fear pro- 
ceedings against them in turn if the University of California were to lose the 
court case or simply, that the project came to a standstill. Perhaps this was the 
reason why later, the installed base of GNU/Linux was much greater than all 
the *BSDs combined. But we cannot know this for sure. 

2.3.3. GNU/Linux comes onstage 

In July 1991 Linus Torvalds (a Finnish 21-year old student) placed his first 
message mentioning his project (at the time) to build a free system similar 
to Minix. In September he released the very first version (0.01), and every 
few weeks new versions would appear. In March 1994 version 1.0 appeared, 
the first one to be called stable, though the kernel that Linus built had been 
usable for several months. During this period, literally hundreds of developers 
turned to Linux, integrating all the GNU software around it, as well as XFree 
and many more free programs. Unlike the *BSDs, the Linux kernel and a large 
number of the components integrated around it were distributed with the GPL 


The story about Linux is probably one of the most interesting (and well-known) in the 
world of free software. You can find many links to information on it from the pages mark- 
ing the 10* anniversary of its announcement, although probably one of the most inter- 
esting ones is the "History of Linux", by Ragib Hasan [138]. As a curiosity you can con- 

GNUFDL' PID 00148386 


Free Software 

suit the thread on which Linus Torvalds announced that he was starting to create what 
later became Linux (in the newsgroup comp.os.minix) at 
groups?th=dl61e94858c4c0b9. There he explains how he has been working on his kernel 
since April and how he has already ported some GNU project tools onto it (specifically 
mentioning Bash and GCC). 

Of the many developments to have emerged around Linux, one of the most 
interesting is the distribution concept 1 . The first distributions appeared soon, 
in 1992 (MCC Interim Linux, of the University of Manchester; TAMU, of Texas 
A&M, and the most well-known, SLS, which later gave rise to Slackware, which 
is still being distributed in the first decade of 2000), entailing the arrival of 
competition into the world of systems packaged around Linux. Each distribu- 
tion tries to offer a ready-to-use GNU/Linux, and starting from the basis of 
the same software has to compete by making improvements considered im- 
portant by their user base. In addition to providing pre-compiled ready-to-use 
packages, the distributions also tend to offer their own tools for managing the 
selection, installation, replacement and uninstallation of these packages, in 
addition to the initial installation on the computer, and the management and 
administration of the operating system. 

( H"his concept is explained in de- 
tail in the corresponding entry in 

Over time, distributions have succeeded each other as different ones became 
the most popular. Of them all, we would highlight the following: 

1) Debian, developed by a community of volunteer users. 

2) Red Hat Linux, which was first developed internally by the company Red 
Hat, but which later adopted a more community-based model, giving rise 
to Fedora Core. 

3) Suse, which gave rise to OpenSUSE, following a similar evolution to Red 

4) Mandriva (successor of Mandrake Linux and Conectiva). 

5) Ubuntu, derived from Debian and produced on the basis of Debian by the 
company Canonical. 

2.4. A time of maturation 

Midway through the first decade of 2000, GNU/Linux, or Fire- 
fox were present in the media quite often. The overwhelming majority of com- 
panies use free software for at least some of their IT processes. It is difficult to 
be an IT student and not to use large amounts of free software. Free software is 
no longer a footnote in the history of IT and has become something very im- 
portant for the sector. IT companies, companies in the secondary sector (those 
that use software intensively, even though their primary activity is different) 

GNUFDL» PID_001 48386 28 Free Software 

and public administrations are starting to consider it as something strategic. 
And slowly but surely it is arriving among domestic users. In broad terms, we 
are entering a period of maturation. 

And at the bottom of it all, an important question starts to arise, which sum- 
marises in a way what is happening: "are we facing a new model of software 
industry?". Perhaps, it may yet happen that free software becomes no more 
than a passing trend to be remembered nostalgically one day. But it may also 
be (and this seems increasingly likely) a new model that is here to stay, and 
perhaps to change radically one of the youngest but also most influential in- 
dustries of our time. 

2.4.1. End of the nineties 

In the mid-1990s, free software already offered complete environments (dis- 
tributions of GNU/Linux, *BSD systems...) that supported the daily work of 
many people, especially software developers. There were still many pending 
assignments (the main one to have better graphical user interfaces at a time 
when Windows 95 was considered the standard), but there were already sev- 
eral thousand people worldwide who used exclusively free software for their 
day to day work. New projects were announced in rapid succession and free 
software embarked on its long path towards companies, the media and public 
awareness in general. 

This period is also associated with Internet taking off as a network for every- 
one, in many cases led by the hand of free programs (especially in its infras- 
tructure). The net's arrival into the homes of millions of end users consolidat- 
ed this situation, at least in terms of servers: the most popular web (HTTP) 
servers have always been free (first the NCSA server, followed by Apache). 

Perhaps the beginning of the road for free software until its full release among 
the public is best described in the renowned essay by Eric Raymond, "The 
cathedral and the bazaar" (Eric S. Raymond, 2001) [192]. Although much of 
what is described in it was already well known by the community of free soft- 
ware developers, putting it into paper and distributing it extensively made it 
an influential tool for promoting the concept of free software as an alternative 
development mechanism to the one used by the traditional software industry. 
Another important paper of this period was "Setting up shop. The Business 
of open source software" [141], by Frank Hecker, which for the first time de- 
scribed the potential business models for free software, and which was written 
in order to influence the decision to release the Netscape Navigator code. 

Whereas Raymond's paper was a great tool for promoting some of the fun- 
damental characteristics of free software, the release of Netscape Navigator's 
code was the first case in which a relatively large company, in a very innova- 
tive sector (the then nascent web industry) made the decision to release one 
of its products as free software. At that time, Netscape Navigator was losing 

GNUFDL» PID_001 48386 29 Free Software 

the browser war against Microsoft's product (Internet Explorer), partly due to 
Microsoft's tactics of combining it with its operating system. Many people be- 
lieve that Netscape did the only thing that it could have done: to try to change 
the rules to be able to compete with a giant. And from this change in the 
rules (trying to compete with a free software model) the Mozilla project was 
born. This project, which had its own problems, has led several years later to a 
navigator that, although it has not recovered the enormous market share that 
Netscape had in its day, seems technically at least as good as its proprietary 

In any case, irrespective of its later success, Netscape's announcement that it 
would release its navigator's source code had a great impact on the software 
industry. Many companies started to consider free software worthy of consid- 

The financial markets also started paying attention to free software. In the 
euphoria of the dotcom boom, many free software companies became targets 
for investors. Perhaps the most renowned case is that of Red Hat, one of the 
first companies to realise that selling CDs with ready-to-use GNU/Linux sys- 
tems could be a potential business model. Red Hat started distributing its Red 
Hat Linux, with huge emphasis (at least for what was common at the time) 
on the system's ease of use and ease of maintenance for people without a spe- 
cific IT background. Over time it diversified, keeping within the orbit of free 
software, and in September 1998 it announced that Intel and Netscape had 
invested in it. "If it is good for Intel and Netscape, it must be good for us", is 
what many investors must have thought then. When Red Hat went public in 
summer 1999, the IPO was subscribed completely and soon the value of each 
share rose spectacularly. It was the first time that a company was obtaining 
financing from the stock exchange with a model based on free software. But 
it was not the only one: later, others such as VA Linux or (which 
was later acquired by VA Linux) did the same. 


Red Hat provides a list of its company milestones at 

During this period, many companies were also born with business models 
based on free software. Despite not going public or achieving such tremendous 
market caps, they were nevertheless very important for the development of 
free software. For example, many companies appeared that started distributing 
their own versions of GNU/Linux, such as SuSE (Germany), Conectiva (Brazil) 
or Mandrake (France), which would later join the former in order to create 
Mandriva. Others offered services to companies that wanted maintenance or 
to adapt free products: LinuxCare (US), Alcove (France), ID Pro (Germany), 
and many more. 

GNUFDL' PID 00148386 


Free Software 

Meanwhile, the sector's giants started to position themselves in relation to 
free software. Some companies, such as IBM, incorporated it directly into their 
strategy. Others, such as Sun Microsystems, had a curious relationship with it, 
at times backing it, at others indifferent, and at others confrontational. Most 
(such as Apple, Oracle, HP, SGI, etc.) explored the free software model with 
various strategies, ranging from the selective freeing of software to straight- 
forward porting of their products to GNU/Linux. Between these two extremes 
there were many other lines of action, such as the more or less intensive use of 
free software in their products (such as the case with Mac OS X) or the explo- 
ration of business models based on the maintenance of free software products. 

From the technical point of view, the most remarkable event of this period was 
probably the appearance of two ambitious projects designed to carry free soft- 
ware to the desktop environment for inexperienced IT users: KDE and GNOME. 
Put simplistically, the final objective was not to have to use the command line 
in order to interact with GNU/Linux or *BSD or with the programs on those 

KDE was announced in October 1996. Using the Qt graphic libraries (at that 
time a proprietary product belonging to the company Trolltech, but free of 
charge for use on GNU/Linux 2 ), construction began of a set of desktop appli- 
cations that would work in an integrated manner and have a uniform appear- 
ance. In July 1998 version 1 .0 of the K Desktop Environment was released, and 
was soon followed by increasingly more complete and more mature new ver- 
sions. GNU/Linux distributions soon incorporated KDE as a desktop for their 
users (or at least as one of the desktop environments that users could choose). 

(2) l_ater, Qt started to be distribut- 
ed under the free licence QPL (Qt 
Public Licence), non-compatible 
with GPL, which caused some 
problems, since most of KDE was 
distributed under the GPL. In time, 
Trolltech finally decided to dis- 
tribute Qt under the GPL licence, 
bringing these problems to an 

Mostly as a reaction to KDE's dependence on the Qt proprietary library, in 
August 1997 the GNOME project was announced (Miguel de Icaza, "The story 
of the GNOME Project") [101], with similar goals and characteristics to those 
of KDE, but stating the explicit objective of all its components being free soft- 
ware. In March 1999, GNOME 1.0 was released, which would also improve 
and stabilise over time. As of that moment, most distributions of free operat- 
ing systems (and many Unix-derived proprietary ones) offered the GNOME or 
KDE desktop as an option, and the applications of both environments. 

Meanwhile, the main free software projects underway remained in good 
health with new projects emerging almost every day. In various niche markets, 
free software was found to be the best solution (acknowledged almost world- 
wide). For example, since its appearance in April 1995, Apache has maintained 
the largest market share for web servers; XFree86, the free project that devel- 
ops X Window, is by far the most popular version of X Window (and therefore, 
the most extended windows system for Unix-type systems); GCC is recognised 
as the most portable C compiler and one of the best quality; GNAT, the com- 
pilation system for Ada 95, has conquered the best part of the market for Ada 
compilers in just a few years; and so on. 

GNUFDL» PID_001 48386 31 Free Software 

In 1998, the Open Source Initiative (OSI) was founded, which decided to 
adopt the term open source software as a brand for introducing free software 
into the business world, while avoiding the ambiguity of the term free (which 
can mean both free to use and free of charge). This decision sparked one 
of the fiercest debates in the world of free software (which continues to 
this day), since the Free Software Foundation and others considered that 
it was much more appropriate to speak about free software (Richard Stall- 
man, "Why free software is better than open source", 1998) [206]. In any case, 
the OSI made a great promotional campaign for its new brand, which has 
been adopted by many as the preferred way to talk about free software, es- 
pecially in the English-speaking world. To define open source software, the 
OSI used a definition derived from the one used by the Debian project to de- 
fine free software ("Debian free software guidelines", 
social_contract.html#guidelines) [104], which at the same time fairly closely 
reflects the idea of the FSF in this regard ("Free software definition", http:// [120], meaning that from the practi- 
cal point of view almost any program considered free software can also be 
considered open source and vice versa. However, the free software and open 
source software communities (or at least the people who identify with them) 
can be very different. 

2.4.2. Decade of 2000 

In the early years of the decade of 2000, free software was already a serious 
competitor in the servers segment and was starting to be ready for the desktop. 
Systems such as GNOME, KDE, and Mozilla Firefox can be 
used by domestic users and are sufficient for the needs of many companies, 
at least where office applications are concerned. Free systems (and especially 
systems based on Linux) are easy to install, and the complexity of maintaining 
and updating them is comparable to that of other proprietary systems. 

Right now, every company in the software industry has a strategy with regards 
to free software. Most of the leading multinationals (IBM, HP, Sun, Novell, 
Apple, Oracle...) incorporate free software to a greater or lesser extent. At one 
extreme we can find companies such as Oracle, which react by simply porting 
their products to GNU/Linux. At another extreme, we can find IBM, which 
has the most decisive strategy and has made the biggest publicity campaigns 
about GNU/Linux. Among the leaders in the IT market, only Microsoft has 
positioned itself in clear opposition to free software and particularly software 
distributed under the GPL licence. 

As regards the world of free software itself, despite the debates that occasion- 
ally stir the community, its growth is massive. Every day there are more devel- 
opers, more active free software projects, more users, etc. With each passing 
day free software is moving away from the sidelines and becoming a force to 
be reckoned with. 

GNUFDL» PID_001 48386 32 Free Software 

In light of this, new disciplines are emerging that specifically study free soft- 
ware, such as free software engineering. Based on research, bit by bit we are 
starting to understand how free software operates in its various aspects: de- 
velopment models, business models, coordination mechanisms, free project 
management, developers' motivations, etc. 

These years we are also starting to see the first effects of the offshoring that free 
software development allows: countries considered "peripheral" are actively 
participating in the world of free software. For example, the number of Mexi- 
can or Spanish developers (both countries with a limited tradition of software 
industry) in projects such as GNOME is significant (Lancashire, "Code, cul- 
ture and cash: the fading altruism of open source development", 2001) [164]. 
And the role of Brazil is even more interesting, with its numerous develop- 
ers and experts in free software technologies, and decisive backing from the 
public administrations. gnuLinEx is a case that merits special attention, as an 
example of how a region with very little tradition of software development 
can try to change the situation through an aggressive strategy of free software 

From the decision-making perspective when it comes to implementing soft- 
ware solutions, we would highlight the fact that there are certain markets 
(such as Internet services or office applications) in which free software is a 
natural choice that cannot be overlooked when studying what type of system 
to use. 

On the negative front, these years have seen how the legal environment in 
which free software operates is changing rapidly worldwide. On the one hand, 
software patents are increasingly adopted in more and more countries. On the 
other hand, new copyright laws make it difficult or impossible to develop free 
applications in some spheres, the most well-known one being DVD viewers 
(due to the CSS encoding algorithm that this technology uses). 


In the beginning of 2002 the Extremadura Regional Government publicly an- 
nounced the gnuLinEx project. The idea was simple: to promote the creation 
of a distribution based on GNU/Linux with the fundamental objective of using 
it on the thousands of computers to be installed in public schools through- 
out the region. Extremadura, situated in the western part of Spain, bordering 
Portugal, has approximately 1 million inhabitants and has never stood out 
for its technological initiatives. In fact, the region had practically no software 

In this context, gnuLinEx has made a very interesting contribution to the free 
software panorama on a global scale. Beyond being just a new distribution of 
GNU/Linux based on Debian (which is still a worthy anecdote), and beyond 
its enormous impact on the mass media (it was the first time that Extremadu- 

GNUFDL» PID_001 48386 33 Free Software 

ra made the front cover of The Washington Post and one of the first that a 
free software product did), what is extraordinary is the (at least apparently) 
solid backing of a public administration for free software. The Regional Gov- 
ernment of Extremadura decided to try a different model where educational 
software was concerned, and then to extend this model to all the software 
used within the scope of its influence. This has made it the first public ad- 
ministration of a developed country to have decisively adopted this approach. 
A lot of interest was generated around the Regional Government's initiative, 
within Extremadura and outside of it: there are academies that teach IT using 
gnuLinEx; books have been written to support this teaching; computers are 
being sold with gnuLinEx pre-installed. In general, they are trying to create 
an educational and business fabric around this experience in order to give it 

support. And the experience has been exported. At the beginning of the 21 st 
century, several autonomous communities in Spain have backed free software 
in education (in one way or another), and in general, its importance for public 
administrations is widely acknowledged. 


Since the end of the nineties, there are GNU/Linux distributions that can be 
easily installed, but Knoppix, whose first version appeared in 2002, has proba- 
bly allowed this idea to reach its full expression. It is a CD that boots on almost 
any PC, converting it (without even having to format the disk, since it can be 
used "live") into a fully functional GNU/Linux machine, with a selection of the 
most frequent tools. Knoppix combines good automatic hardware detection 
with a good choice of programs and "live" functioning. For example, it allows 
a rapid and direct experience of what it means to work with GNU/Linux. And 
it is giving rise to an entire family of distributions of the same type, specialised 
for the specific requirements of a user profile. 

In 1999, Sun Microsystems bought a German company called Stardivision, 
whose star product was StarOffice, a suite of office applications similar in func- 
tionality to the Microsoft Office set of tools. One year later, Sun distribut- 
ed most of the StarOffice code under a free licence (the GPL) creating the project. This project released version 1.0 of 
in May 2002. has become a quality suite of office applica- 
tions with a similar functionality to that of any other office product, and, 
more importantly, it interoperates very well with the Microsoft Office data 
formats. These features have made it the reference free software application 
in the world of office suites. 

The importance of, from the point of view of extending free 
software to a large number of users, is enormous. Finally it is possible to 
change, almost without problems, from the proprietary environments com- 
mon with office suites (undoubtedly the star application in the business 

GNUFDL' PID 00148386 


Free Software 

world) to totally free environments (such as GNU/Linux plus GNOME and/or 
KDE plus Also, the transition can be made very smoothly: 
since also works on Microsoft Windows, it is not necessary 
to change operating systems in order to experiment in depth with using free 

M ozilla, Firefox and the rest 

Practically since its appearance in 1994 until 1996, Netscape Navigator was 
the unchallenged market leader in web browsers, with market shares of up 
to 80%. The situation started to change when Microsoft included Internet Ex- 
plorer with Windows 95, causing Netscape Navigator to gradually lose mar- 
ket share. At the beginning of 1998 Netscape announced that it was going to 
distribute a large part of its navigator code as free software, which it did in 
March that same year, launching the Mozilla project. For quite a while the 
project was clouded by uncertainty, and even pessimism (for example, when 
its leader, Jamie Zawinski, abandoned it), because as time went by no product 
was resulting from its launch. 

In January 2000, the project released Mozilla M13, which was considered the 
first relatively stable version. In May 2002 version 1.0 was finally published, 
the first officially stable version, over four years after the first Netscape Navi- 
gator code had been released. 

Finally Mozilla had become a reality, although perhaps too late, if we bear in 
mind the market shares that Internet Explorer had in 2002 or 2003 (when it 
was the undisputed leader leaving Mozilla and others in a totally marginal 
position). But despite taking so long, the Mozilla project has borne fruit; not 
only expected fruit (the Mozilla navigator), but also other "collateral" ones, 
such as Firefox for example, another navigator based on the same HTML en- 
gine, which has become the main product, and which since it appeared in 
2005 is managing bit by bit to erode other navigators' market share. 


In "Netscape Navigator", by 
Brian Wilson, [234], we can 
consult a detailed list of the 
main versions of Netscape 
Navigator and Mozilla, and 
their main characteristics. 

The Mozilla project has helped to fill a large gap in the world of free soft- 
ware. Before Konqueror appeared (the KDE project's navigator), there were not 
many free navigators with a graphic interface. Since the publication of Mozil- 
la, an enormous number of projects based on it have emerged which have 
produced a large number of navigators. At the same time, the combination 
of Mozilla Firefox and allows free software to be used for the 
most common tasks, even in a Microsoft Windows environment (they both 
work not only on GNU/Linux, *BSD and other Unix-type systems, but also on 
Windows). For the first time in the history of free software, it has made the 
transition from proprietary software to free software in office environments a 
simple task: we can start by using these two applications on Windows, without 
changing operating systems (for those who use it normally), and over time 
eliminate the only non-free part and move onto GNU/Linux or FreeBSD. 

GNUFDL" PID_001 48386 35 Free Software 

The case of SCO 

At the beginning of 2003, the SCO corporation (formerly Caldera Systems and 
Caldera International) presented a legal case against IBM for alleged breach 
of its intellectual property rights. Although the case was complex, it centred 
on the accusation that IBM had contributed to the Linux kernel with code 
belonging to SCO. In May 2007, the matter had still not been resolved and had 
even become more complicated by further legal suits (IBM and Red Hat against 
SCO, SCO against AutoZone and DaimlerChrysler, two large IT users) and by 
SCO's campaigns threatening to pursue big companies that used Linux, etc. 

Although the winner of this enormous legal battle has still not emerged, the 
case has highlighted certain legal aspects concerning free software. In particu- 
lar, many companies have considered the problems that they may have to face 
if they use Linux and other free programs, and the guarantee that in doing so 
they are not in breach of third party intellectual or industrial property rights. 

In some way, this case and other ones (such as those related to the validity of 
the GPL licences which were resolved in Germany in 2005) may also be inter- 
preted as a sign of the maturity of free software. It has stopped being a stranger 
to the business world to become part of many of its activities (including those 
related to legal strategies). 

Ubuntu, Canonical, Fedora and Red Hat 

Although Canonical (the company that produces and distributes Ubuntu) 
could be considered a recent arrival to the business of GNU/Linux distribu- 
tions, its activities deserve our attention. In a relatively short time, Ubuntu 
has established itself as one of the best known and most widely used distribu- 
tions, with a reputation for good quality, and great ease of installation and 
use. Ubuntu also stands out for its greater attention to including fundamen- 
tally free software than most distributions produced by companies. 

However, the fundamental characteristic of Ubuntu (and of Canonical's strat- 
egy) has been to base on Debian, a distribution created and maintained by 
volunteers. In fact, Ubuntu is not the first case of a distribution based on De- 
bian (another well-known case is gnuLinEx), but perhaps it is the one to have 
received the most funding. For example, Canonical has hired a large number 
of Debian experts (many of whom participate in the project) and has pursued 
a strategy that seeks collaboration with the volunteer project. To some extent, 
Canonical has tried to fill what it considers is missing from Debian in order 
to gain acceptance from the average user. 

Red Hat, in turn, has followed a different path in order to wind up in a fairly 
similar situation. Starting from a distribution produced entirely with its own 
resources, it decided to collaborate with Fedora, a group of volunteers that 
was already working with distributions based on Red Hat, in order to produce 

GNUFDL" PID_001 48386 36 Free Software 

Fedora Core, its "popular" distribution. Red Hat maintains its version for com- 
panies, but this collaboration with volunteers is, in the end, very similar to 
the one that has produced Ubuntu. 

Perhaps all of these movements are no more than the product of the fierce 
competition taking place in the market for GNU/Linux distributions and of 
one more notable trend: companies' collaboration with volunteers (with the 
community) to produce free software. 

Customised distributions 

Since Linux came onto the scene, a large number of groups and companies 
have created their own distributions based on it. But during these years, the 
phenomenon has caught on with many organisations and companies that 
want customised versions for their own requirements. Customisation has 
been able to expand because the process has become cheaper and there is 
widespread availability of the technical knowledge to do so, even making this 
a niche market for certain companies. 

Perhaps one of the best known cases of customised distributions is the one for 
Spain's autonomous communities. The Extremadura Regional Government 
with its gnuLinEx sparked a trend that many other autonomous communities 
have since followed. The process is so common that several of them regularly 
convene tenders for the creation and maintenance of new versions of their 

The creation of customised distributions realises a trend that the world of free 
software had been discussing for a long time: adapting programs to users' spe- 
cific needs without it having to be the original producers that necessarily make 
the adaptation. 


Some of the most well-known distributions of GNU/Linux in the Spanish autonomous 
communities include: 

• gnuLinEx: (Extremadura) 

• Guadalinex: (Andalucia) 

• Lliurex: (Comunidad Valenciana) 

• Augustux: (Aragon) 

• MAX: (Madrid) 

• MoLinux: (Castilla-La Mancha) 

Company-company and volunteer-company collaborations 

Since practically the beginning of free software, there have been companies 
that collaborated with volunteers in developing applications. However, in 
these years when it appears that we are reaching maturity there is a growing 

GNUFDL» PID_001 48386 37 Free Software 

number of companies that use free software as part of their strategy to collab- 
orate with other companies, when they find it interesting. Two of the most 
significant cases, organised specifically with this objective, are ObjectWeb (an 
alliance formed in France which over time clearly has clearly become interna- 
tional) and Morfeo (in Spain). In both cases, a group of companies has agreed 
to develop a set of free systems that are of interest to them, and decided to 
distribute it as free software. 

In other cases, companies have actively sought to collaborate in free projects 
promoted by volunteers, or tried to make volunteers collaborate with 
their own free projects. The GNOME Foundation or the already-mentioned 
Ubuntu in respect of Debian are examples of this first scenario. Sun and and OpenSolaris, or Red Hat with Fedora Core, are examples 
of the second. 

Expanding to other spheres 

Free software has proven that in the field of producing programs there is an- 
other way of doing things. In practice, we have seen how granting the freedom 
to distribute, modify and use can achieve sustainability, either through volun- 
teer work, or through business generation that allows companies to survive. 

As time passes, this same idea is being transferred to other spheres of intel- 
lectual work. The Creative Commons licences have made it possible to free 
spheres such as literature, music, or video. Wikipedia is proving that a field as 
particular as the production of encyclopaedias can move through a very inter- 
esting path. And there are more and more literary authors, music bands and 
even film producers interested in models of free production and distribution. 

In all these domains there is still a long way to go, and in almost all of them 
practice has not yet fully proven that sustainable creation is possible with free 
models. But it cannot be denied that experimentation with it is reaching a 
boiling point. 

Free software as a subject of study 

Although some works, such as the renowned "The cathedral and the bazaar" 
cleared the way for the study of free software as such, it was not until 2001 and 
subsequent years that the academic community started to consider free soft- 
ware as something worthy of study. Over time, the massive availability of data 
(almost everything in the world of free software is public and available from 
public information archives) and the innovations that free software has pro- 
vided have drawn the attention of many groups. Midway through the decade 
of 2000 there are already several international conferences centred specifically 
on free software, top-ranking magazines frequently produce papers on it, and 
research-funding agencies are opening lines aimed specifically towards it. 

GNUFDL" PID_001 48386 38 Free Software 

2.5. The future: an obstacle course? 

Of course, it is difficult to predict the future. And that is certainly not our ob- 
jective. Therefore, rather than trying to explain what the future of free soft- 
ware will be like, we will try to show the problems that it will foreseeably have 
to face (and has indeed been facing for a long time). How the world of free 
software is able to overcome these obstacles will undoubtedly determine its 
situation in several years' time. 

• FUD (fear, uncertainty, doubt). This is a fairly common technique in the 
world of information technologies, used by free software's competitors in 
order to discredit free software, with more or less justification and varying 
degrees of success. In general terms, free software has been fairly immune 
to these techniques, perhaps due to its complexity and different ways of 
seeping into companies. 

• Dissolution. Many companies are testing the limits of free software as 
a model, and in particular are trying to offer their clients models that 
present some similar characteristics to free software. The main problem 
that can present itself with this type of model is that it generates confu- 
sion among clients and developers, who need to read the small print in 
detail in order to realise that what they are being offered does not have the 
advantages that free software offers them. The most well-known model of 
this type is the Shared Source program, by Microsoft. 

• Lack of knowledge. In many cases, users turn to free software simply be- 
cause they think that it is free of charge; or because they think that it is 
"fashionable". If they do not look deeper into it, and study with a certain 
amount of detail the advantages that free software can offer as a model, 
they run the risk of not taking full advantage of them. In many cases, 
the initial assumptions in the world of free software are so different from 
the traditional ones in the world of proprietary software that a minimum 
analysis is required in order to understand that what in one case is fre- 
quent in the other may be impossible, and vice versa. Therefore, lack of 
knowledge can only generate dissatisfaction and loss of opportunities for 
any person or organisation approaching free software. 

• Legal obstacles. This is certainly the main problem that free software is go- 
ing to have to deal with in coming years. Although the legal environment 
in which free software developed in the 80s and first half of the 90s was 
not ideal, at least it left enough space for it to grow freely. Since then, ex- 
tension of the scope of patenting to software (which has occurred in many 
developed countries) and new copyright legislation (limiting the software 
developer's liberty to create) are producing increasingly higher barriers to 
free software's entry into important segments of applications. 

GNUFDL' PID 00148386 


Free Software 

2.6. Summary 

This chapter presents the history of free software. The sixties was a period 
dominated by large computers and IBM in which software was distributed 
together with the hardware, and usually with the source code. In the seven- 
ties, software started to be sold separately, and soon proprietary distributions, 
which did not include source code and did not give permission to modify or 
redistribute, became almost the only option. 

See also 

Interested readers will find in 
Appendix B a list of some of 
the most relevant dates in the 
history of free software. 

In the decade of the 1970s, work began on developing the Unix operating 
system at AT&T's Bell Labs, giving rise later to Unix BSD. Its evolution, in 
parallel with the birth of the Internet, served as a testing field for new ways 
of developing in collaboration, which later became common in the world of 
free software. 

In 1984, Richard Stallman started to work on the GNU project, founding the 
Free Software Foundation (FSF), writing the GPL licence, and in general estab- 
lishing the foundations of free software as we now know it. 

In the 90s Internet matured offering free software communities new channels 
for communication and distribution. In 1991, Linus Torvalds started to devel- 
op a free kernel (Linux) which helped to complete the GNU system, which 
already had almost all the parts for becoming a complete system similar to 
Unix: C compiler (GCC), editor (Emacs), windowing system (X Window), etc. 
This is how the GNU/Linux operating systems were born, branching out into 
many distributions, such as Red Hat Linux and Debian GNU/Linux. Towards 
the end of the 90s, these systems were completed with two desktop environ- 
ments: KDE and GNOME. 

In the decade of 2000, free software managed to lead in some sectors (such 
as for web servers, dominated by Apache), and new tools appeared covering 
a large number of IT requirements. 

GNUFDL' PID 00148386 


Free Software 

3. Legal aspects 

"The licences for most software are designed to take away your freedom to share and 
change it." 

GNU General Public Licence, version 2 

This chapter looks at the main legal aspects related to free software. To put 
them into context, we start with a small introduction to the most basic con- 
cepts of intellectual and industrial property rights, before offering the detailed 
definition of free software, open source software and other related concepts. We 
also look in some detail at the most common free software licences and their 
impact on business models (subject covered in greater detail in chapter 5) and 
development models. 

3.1. Brief introduction to intellectual property 

The term intellectual property has various meanings according to its context 
and who uses it. Nowadays it is frequently used in many spheres to refer to 
various privileges awarded over intangible goods with economic value. It in- 
cludes concepts such as copyright and similar, which protect from unautho- 
rised copy literary or artistic works, computer programs, data compilations, 
industrial designs, etc.; trademarks, which protect symbols; geographical in- 
dications, which protect appellations of origin; trade secrets, which protect 
the hiding of information, and patents, which concede temporary monopo- 
lies to inventions in exchange for their revelation. However, in many legal 
traditions, including the Hispanic tradition, a distinction is made between in- 
tellectual property, which refers exclusively to copyright, and industrial property, 
which covers the other concepts. 

In any case, the legislation applicable to all of these aspects is one of the best 
coordinated practically worldwide. On the one hand, the WIPO (Worldwide 
International Property Organisation) covers both types of property in all of 
their aspects. On the other hand, the TRIPS agreement (Trade-Related aspects 
of Intellectual Property rights) establishes certain minimum levels of protec- 
tion and obliges all member countries of the WTO (World Trade Organisation) 
to develop them within certain timeframes, according to the level of develop- 
ment of the country 3 

( H"he TRIPS agreement was signed 
under pressure from the industri- 
alised countries (especially the US 
and Japan). 

Article 27 of the Declaration of Human Rights acknowledges that everyone 
has the right to the protection of the moral and material interests resulting 
from any scientific, literary or artistic production of which he is the author. 
However, in many cases (and frequently in the case of software), this right 
is transferred in practice to the companies that employ the creators or that 
distribute or sell their creations. Nonetheless, intellectual property is justified 

GNUFDL» PID_001 48386 41 Free Software 

not just morally, but also for practical reasons, in order to comply with an- 
other right: the public's right to benefit from creation, promoting it through 
incentives and protecting investments in creation, research and development. 
In order to harmonise these two rights, intellectual property is temporary and 
expires once it has fulfilled its function of promotion. 

But expiry is not the only distinguishing feature between intellectual prop- 
erty and ordinary property. Nowadays, its products can be copied easily and 
cheaply, without any loss of quality. Copying does not prejudice the party that 
is already benefiting from what is copied, unlike theft, which does deprive 
the original possessor. Copying can prejudice the owner, by depriving him of 
potential income from a sale. Controlling the copying of intangibles is much 
more complicated than controlling the theft of tangible property and can lead 
us to a situation of a police state, having to control all copies of information, 
and legal insecurity since the potential for accidental infringement of rights 
increases. Furthermore creativity is incremental: creating always copies some- 
thing, and the dividing line between a poor imitation and inspiration is a 
subtle one. 

In order to study this in more depth, the following sections go over some of 
the categories of intellectual property. In any case, we can already advance that 
free software proposes a new point of equilibrium in this sphere, advocating 
the benefits of copying and incremental innovation versus exclusive control 
of a work by its author. 

3.1.1. Copyright 

Copyright protects the expression of a content, not the content itself. Copy- 
right was developed in order to compensate the authors of books or art. Pro- 
tected works may express ideas, knowledge or methods that are freely usable, 
but it is prohibited to reproduce them without full or partial permission, with 
or without modifications. This protection is very simple, since it automatical- 
ly comes into force with an almost universal scope just when the work is pub- 
lished/released. Currently, it has been extended to computer programs and (in 
some geographical areas) to data compilations. 

The Law on Intellectual Property (LPI) in Spain, and similar laws in other coun- 
tries, developed on the basis of the Berne Convention of 1886 for the protec- 
tion of literary and artistic works, regulates copyright. These rights are divided 
into moral and intellectual rights. The former guarantee the author's control 
over the distribution of his work, under his name or pseudonym, the recogni- 
tion of authorship, respect for the integrity of the work and the right to mod- 
ify and withdraw it. The second give the author the right to exploit the work 
economically and may be ceded in whole or in part, exclusively or not, to a 

GNUFDL» PID_001 48386 42 Free Software 

third party. Moral rights are lifelong or indefinite, whereas intellectual rights 
have a fairly long duration (seventy years following the author's death, in the 
case of a physical person and Spanish law). 

Cession of these rights is established by means of a contract known as a 
licence. In the case of proprietary programs, these are generally distributed 
through "non exclusive" licences for use, understood as automatically accept- 
ed by opening or installing the product. Therefore it is not necessary to sign 
the contract, since in the case of the receiver not accepting it, the rights by 
default under the law govern automatically, that is none. Licences cannot re- 
strict some of the rights granted by current legislation, such as the right to 
make private copies of art or music, which allows a copy of a recording to be 
given to a friend as a gift, but this right does not apply to programs. According 
to the LPI of 1996 (Spanish Law on Intellectual Property. Royal Legislative De- 
cree 1/1996, of 12th April) [77], modified in 2006 (Law on Intellectual Prop- 
erty. Law 23/2006, of 7 l July) [79], in respect of programs it is always possible 
to make a backup copy, they may be studied to be made interoperable and 
they may be corrected and adapted to our needs (which is difficult, because 
normally we do not have access to the source code). These rights can not be 
restricted through licences, although the laws are under review, in an appar- 
ently unstoppable trend to limit the rights of users. Organised compilations 
of works or third party data are also subject to copyright, although under dif- 
ferent terms with a shorter timeframe. 

New information technologies, and specially the web, have deeply trans- 
formed copyright protection, since expressions of content are much easier to 
copy than content itself. And in the case of programs and some works of art 
(music, images, films, and even literature) they "work" automatically on the 
computer without the need for any appreciable human effort. However, de- 
signs or inventions need to be built and possibly put into production. This 
possibility of generating wealth at no cost has led a large proportion of the 
public, in particular in poor countries, to duplicate programs without paying 
the licence, without public awareness of this being a "malicious action" (unlike 
in the case of stealing physical property, for example). Meanwhile, program 
manufacturers, either alone or in coalition (through the BSA, Business Soft- 
ware Alliance, for example), exert enormous pressure for licences to be paid 
and for governments to pursue what has become known as piracy. 


The word piracy has become generally accepted as a synonym for the 'violation of any 
form of intellectual property, especially in the case of illegally copying of programs, mu- 
sic and films'. The term seems exaggerated and in the dictionary of the Royal Spanish 
Academy of Language it appears with that meaning in the figurative sense, since the 
original word refers to 'robbery with violence committed at sea'. This is why Richard 
Stallman recommends avoiding it ("Some confusing or loaded words and phrases that 
are worth avoiding", 2003) [212]. 

GNUFDL» PID_001 48386 43 Free Software 

Precisely in order to protect the copyright of contents with proprietary li- 
cences, the so-called DRM systems were born (digital rights management), de- 
signed to control access and the use of data in digital format or to restrict its 
use to certain devices. The use of DRM systems has been strongly criticised 
in many sectors, because they protect copyright imposing restrictions beyond 
what is sufficient, which is why some, such as the Free Software Foundation, 
recommend interpreting the acronym as digital restrictions management, in an 
attempt to avoid using the word rights, because it considers that there is an 
excessive deprivation of the rights of users in favour of satisfying copyright 

3.1.2. Trade secret 

Another resource that companies use in order to make profit from their invest- 
ments is trade secret, protected by the laws of industrial property, provided 
that companies take sufficient measures to hide the information that they do 
not wish to disclose. In the case of chemical or pharmaceutical products that 
require governmental approval, the government undertakes not to disclose 
submitted data that is not mandatory to make public. 

One of the best known applications of trade secret is the propietary software 
industry, which generally sells compiled programs without access to the source 
code, in order to prevent derived programs from being developed. 

At first sight it would appear that the protection of trade secret is perverse, 
since it can deprive society of useful knowledge indefinitely. To some extent 
some legislations also interpret it this way, and allow reverse engineering in 
order to develop replacement products, although industry pressure has man- 
aged to prohibit this activity in many countries, and in other countries only 
made it possible on the grounds of compatibility. 

Whether or not trade secret is a perversion, in many cases it is better than a 
patent since it offers a competitive advantage to the person placing a product 
on the market while the competition tries to imitate it through reverse engi- 
neering. The more sophisticated the product the more it will cost the compe- 
tition to reproduce it, whereas if it is trivial, it will be copied quickly. Imita- 
tion with improvements has been fundamental in the development of today's 
superpowers (the US and Japan) and is very important for the financial inde- 
pendence of developing countries. 

3.1.3. Patents and utility models 

The alternative to trade secret is a patent. In exchange for a seventeen to twen- 
ty five year monopoly and a specific financial cost, an invention is publicly 
disclosed so that it can be easily reproduced. It aims to promote private re- 

GNUFDL" PID_001 48386 44 Free Software 

search, at no cost to the taxpayer and without losing the outcome. The patent 
holder can decide whether to allow others to use it and the price to be paid 
for the licence. 

Official doctrine is that the patent system promotes innovation, but more 
and more voices are making themselves heard with the view that it impedes 
it, either because the system is poorly implemented or because they consider 
that it is perverse in itself (Francois-Rene Rideau, "Patents are an economic 
absurdity", 2000) [194]. 

What is considered an invention has changed over time, and there is enor- 
mous pressure to extend the scope of the system, to include algorithms, pro- 
grams, business models, natural substances, genes and forms of life, includ- 
ing plants and animals. TRIPS requires the patents system not to discriminate 
against any field of knowledge. The pressures of the World Intellectual Proper- 
ty Organisation (WIPO) aim to eliminate the need for an invention to have an 
industrial application and also to reduce the standards of invention required 
of a patent. The US is at the forefront of countries with minimum require- 
ments on patentability, and is also the most belligerent for other countries to 
adopt its standards, forgetting that the US refused to accept foreign patents 
when it was an underdeveloped country. 

After obtaining a patent, the rights of the owner are independent of the quality 
of the invention and the effort invested in obtaining it. Given the cost of 
maintaining a patent, and litigation costs, only large companies are able to 
maintain and do maintain a large portfolio of patents, which puts them in 
a position to strangle any competition. Given the ease of getting patents on 
trivial or widely applicable solutions, they can thus monopolise an extensive 
field of economic activity. 

With patents, many activities, especially programming, become extremely 
risky, because it is very easy that in developing a complicated program there 
is an accidental violation of some patent. When two or more companies are 
conducting research in order to resolve a problem, it is highly probable that 
they will reach a similar solution at almost the same time, but only one of 
them (usually the one with most resources) will manage to patent its inven- 
tion, preventing the others from having any chance of recouping their invest- 
ment. Any complex technological development becomes a nightmare if in or- 
der to solve each part you first need to find out whether the solution found 
is patented (or patent pending), so as to obtain the licence or find an alterna- 
tive solution. This problem is particularly severe with free software, where the 
violation of algorithm patents is evident from simply inspecting the code. 

Although in Europe it is still illegal to patent an algorithm, it will become 
possible in the near future, perhaps by the time the reader reads these lines. 

GNUFDL» PID_001 48386 45 Free Software 

3.1.4. Registered trademarks and logos 

Trademarks and logos are names and symbols that represent an established 
quality (or a massive investment in publicity). They are not very important 
in the world of free software, possibly because registering them has a cost. 
Therefore, just a few important names such as Open Source (by the Open 
Source Foundation), Debian (by Software in the Public Interest), GNOME 
(by the GNOME Foundation), GNU (by the Free Software Foundation) or (by SUN Microsystems) are registered, and only in a few coun- 
tries. However, not registering the names has caused problems. For example, 
in the US (1996) and in Korea (1997) people have registered the name Linux 
and demanded payment for its use. Resolving these disputes entails legal costs 
and the need to prove the use of the name prior to the date of registration. 

3.2. Free software licences 

Legally speaking, the situation of free programs in relation to proprietary ones 
is not very different: they are both distributed under a licence. The difference 
lies in what the licence allows. In the case of free program licences, which do 
not restrict particularly their use, redistribution and modification, what can be 
imposed are conditions that need to be met precisely in the case of wanting to 
redistribute the program. For example, it is possible to demand observation of 
authorship indications or to include the source code if wanting to redistribute 
the program ready to run. 

Although essentially free software and prop ietary software differ in terms of the 
licence under which the authors publish their programs, it is important to 
emphasise that this distinction is reflected in completely different conditions 
of use and redistribution. As we have seen in the last few years, this has not 
only given rise to totally different development methods, but also to practi- 
cally opposite ways (in many aspects) of understanding IT. 

The laws on intellectual property ensure that in the absence of explicit per- 
mission virtually nothing can be done with a work (in our case, a program) 
received or purchased. Only the author (or the holder of the rights of the 
work) can grant us that permission. In any case, ownership of the work does 
not change by granting a licence, since this does not entail transfer of own- 
ership, but rather just the right of use, and in some cases (mandatory with 
free software), of distribution and modification. Free software licences are dif- 
ferent from proprietary software licences precisely in that instead of carefully 
restricting what is allowed, it makes certain explicit allowances. When people 
receive a free program they may redistribute it or not, but if they do redis- 
tribute it, they can only do so because the licence allows it. But to do so the 
licence must be observed. Indeed, the licence contains the rules of use that 
users, distributors, integrators and all other parties involved in the world of 
IT must observe. 

GNUFDL» PID_001 48386 46 Free Software 

In order to fully understand all the legal ins and outs that arise in this chapter 
(and which are without question very important to understand the nature of 
free software) we should also be aware that each new version of a program is 
considered a new work. The author, once again, is fully entitled to do what he 
wants with it, even to distribute it under totally different terms and conditions 
(in other words, with a totally different licence to the earlier one). That way if 
the reader is the sole author of a program, he may publish one version under 
a free software licence and, if he wishes to, a later one under a proprietary li- 
cence. In the case of there being more authors and the new version containing 
code that they have produced, if the code is to be published under different 
conditions, all of them will have to approve the change in licence. 

A still relatively open issue is the licence that applies to external contributions. 
Generally it is assumed that someone who contributes to a project accepts 
de facto that his or her contribution adjusts to the conditions specified by its 
licence, although the legal grounds for this are poor. The initiative of the Free 
Software Foundation to ask by means of a (physical) letter for cession of all 
copyright from anyone who contributes more than ten lines of code to a GNU 
sub-project is a clear indication that in the world of free software there are 
stricter policies with regards to these contributions. 

Based on the foregoing, in the rest of the chapter we will focus on analysing 
different licences. To place this analysis into context, we must remember that 
from now on, when we say that a licence is a free software licence, what we 
mean is that it fulfils the definitions of free software presented in section 1.1.1. 

3.2.1. Types of licences 

There is an enormous variety of free licences, although for practical reasons 
most projects use a small group of four or five. On the one hand, many projects 
don't want to or don't have the resources to design their own licence; on the 
other hand, most users prefer to refer to a well-known licence than having to 
read and analyse complete licences. 


There is a compilation and discussion of the licences considered non-free or free but in- 
compatible with the GPL from the point of view of the FSF in the Free Software Founda- 
tion, "Free licences" [121] . The philosophically different point of view of the Open Source 
Initiative is shown in its list (Open Source Initiative, "Open Source licences") [181]. We 
can see discrepancies in some licences, such as the Apple Public Source Licence Ver. 1.2, 
which the FSF considers non-free because of the obligation to publish all changes (even 
if they are private), to notify Apple of redistributions, or the possibility of unilateral re- 
vocation. Nevertheless, the pressure of this classification made Apple publish its version 
2.0 in August 2003, which the FSF then did consider free. 

GNUFDL' PID 00148386 


Free Software 

It is possible to divide free software licences into two large families. The first 
comprises licences that do not impose special conditions on the second redistri- 
bution (in other words, that only specify that the software can be redistributed 
or modified, but that do not impose special conditions for doing so, which 
allows, for example, someone receiving the program to then redistribute it 
as proprietary software): these are what we will refer to as permissive licences. 
The second family, which we will call strong licences (or copyleft licences), in- 
clude those that, in the style of GNU's GPL, impose conditions in the event of 
wanting to redistribute the software, aimed at ensuring compliance with the 
licence's conditions following the first redistribution. Whereas the first group 
emphasises the freedom of the person receiving the program to do almost 
anything he or she wants with it (in terms of the conditions for future redis- 
tributions), the second emphasises the freedom of anyone who may poten- 
tially receive some day a work derived from the program, obliging subsequent 
modifications and redistributions to respect the terms of the original licence. 


The term copyleft when ap- 
plied to a licence, used main- 
ly by the Free Software Foun- 
dation to refer to its own li- 
cences, has similar implications 
to those referred to as strong li- 
cences as used in this text. 

The difference between these two types of licences has been (and remains) a 
debatable issue amongst the free software community. In any case, we should 
remember that they are all free licences. 

3.2.2. Permissive licences 

Permissive licences, also known sometimes as liberal or minimal licences, do 
not impose virtually any conditions on the person receiving the software, and 
yet, grant permission to use, redistribute and modify. From a certain point of 
view, this approach can be seen as a guarantee of maximum freedom for the 
person receiving the program. But from another, it may also be understood 
as maximum neglect in respect of ensuring that once someone receives the 
program, that person guarantees the same freedoms when redistributing that 
program. In practice, these licences typically allow software that its author 
distributes under a permissive licence to be redistributed with a proprietary 

Among these licences, the BSD licence is the best known, to such an extent 
that often permissive licences are referred to as BSD-type licences. The BSD 
(Berkeley Software Distribution) licence stems from the publication of differ- 
ent versions of Unix produced by the University of California in Berkeley, in 
the US. The only obligation it imposes is to credit the authors, while it allows 
redistribution in both binary and source code formats, without enforcing ei- 
ther of the two in any case. It also gives permission to make any changes and 
to be integrated into other programs without almost any restrictions. 

GNUFDL» PID_001 48386 48 Free Software 


One of the consequences in practice of BSD-type licences has been to diffuse standards, 
since developers find no obstacle to making programs compatible with a reference im- 
plementation under this type of licence. In fact, this is one of the reasons for the ex- 
traordinary and rapid diffusion of Internet protocols and the sodcefs-based programming 
interface, because most commercial developers derived their implementation from the 
Berkeley University one. 

Permissive licences are fairly popular, and there is an entire family with sim- 
ilar characteristics to the BSD: X Window, Tcl/Tk, Apache, etc. Historically, 
these licences appeared because the corresponding software was developed by 
universities with research projects financed by the US Government. The uni- 
versities did not sell these programs, on the assumption that they had already 
been paid for by the Government, and therefore by the taxpayer, which meant 
that any company or individual could use the software without almost any 

As already mentioned, on the basis of a program distributed under a permis- 
sive licence another one can be created (in reality, a new version), which may 
be proprietary. Critics of BSD licences see a danger in this feature, because it 
does not guarantee the freedom of future program versions. Its supporters, on 
the contrary, consider it the maximum expression of freedom and argue that, 
ultimately, you can do (almost) everything you want with the software. 

Most permissive licences are a word for word copy of Berkeley's original, mod- 
ifying just what refers to authorship. In some cases, such as the Apache project 
licence, it includes an additional clause, such as prohibiting giving the same 
name as the original to redistributed versions. All of these licences usually in- 
clude, like BSD, the prohibition to use the name of the rightholder for pro- 
moting derived products. 

At the same time, all the licences, whether BSD-type or not, include a limi- 
tation of guarantee which is really a disclaimer, necessary in order to avoid le- 
gal claims for implicit guarantees. Although this disclaimer in free software 
has been broadly criticised, it is common practice with proprietary software, 
which generally only guarantees that the medium is correct and that the pro- 
gram in question runs. 

GNUFDL» PID_001 48386 49 Free Software 

Summary outline of the BSD licence 

Copyright © the owner. All rights reserved. 

Redistribution and use in source and binary forms, with or without modification, are 
permitted provided that the following conditions are met: 

1) Redistributions of source code must reproduce the copyright notice, and list these 
conditions and the disclaimer. 

2) Redistributions in binary form must reproduce the copyright notice and list these 
conditions and the disclaimer in the documentation. 

3) Neither the name of the owner nor the names of its contributors may be used to 
endorse products derived from this software without permission. 

This program is provided "as is", and any express or implicit warranties, of mer- 
chantability or fitness for a particular purpose are disclaimed. In no event shall the 
owner be liable for any damage caused by its use (including loss of data, loss of prof- 
its or business interruption). 

Next we describe in brief a few permissive licences: 

• X Window licence, version 11 (XI 1) 
( [73] . 

This is the licence used to distribute the X Window system, the most ex- 
tensively used windows system in the world of Unix, and also GNU/Lin- 
ux environments. It is very similar to the BSD licence, which allows redis- 
tribution, use and modification without practically any restrictions. It is 
sometimes called the MIT licence (with a dangerous lack of precision, since 
MIT has used other types of licences). Works derived from the X Window 
System, such as XFree86 are also distributed under this licence. 

• Zope Public Licence 2.0 ( [76]. 

This licence (commonly referred to as ZPL) is used for the distribution of 
Zope (an application server) and other related products. It is similar to BSD, 
with the curious feature that it expressly prohibits the use of trademarks 
registered by the Zope Corporation. 

• Apache licence. 

This is the licence under which most of the programs produced by the 
Apache project are distributed. It is similar to the BSD licence. 

There are some free programs that are not distributed with a specific licence, 
rather the author explicitly declares them to be public domain. The main out- 
come of this declaration is that the author waives all rights to the program, 
which can therefore be modified, redistributed, used, etc., in any way. In prac- 
tical terms, it is very similar to a program being under a BSD-type licence. 

GNUFDL» PID_001 48386 50 Free Software 

3.2.3. Strong licences 

The GNU General Public Licence (GNU GPL) 

The General Public Licence of the GNU project (Free Software Foundation, 
1991) [118] (better known by its acronym GPL), which appears in appendix C, 
is by far the most popular and well-known in the world of free software. It was 
created by the Free Software Foundation (promoter of the GNU project), and 
was originally designed to be the licence for all software generated by the FSF. 
However, its use has extended further becoming the most used licence (for 
example, more than 70% of the projects announced on Freshmeat are licensed 
under the GPL), even by flagship projects in the world of free software, such 
as the Linux kernel. 

The GPL licence is interesting from a legal point of view because it makes 
creative use of copyright legislation, to achieve practically the opposite effect 
of what that legislation intends: instead of limiting users rights, it guarantees 
them. For this reason, this manoeuvre is frequently called copyleft (a play on 
words by replacing the word "right" with "left"). Someone with a sense of hu- 
mour even devised the slogan "copyleft, all rights reversed". 

In basic terms, the GPL licence allows redistribution in binary form and in 
source code, although in the case of a binary redistribution access to the source 
code is also obligatory. It also allows modifications to be made without any 
restrictions. However, it is only possible to redistribute code licensed under 
the GPL integrated with other code (for example, linking the code) if it has a 
compatible licence. This has been called the viral effect (although many con- 
sider this name to be disrespectful) of the GPL, since once code has been pub- 
lished with these conditions they can never be changed. 


A licence is incompatible with the GPL when it restricts any of the rights guaranteed by 
the GPL, either explicitly by contradicting any of its clauses, or implicitly, by imposing 
a new limitation. For example, the current BSD licence is compatible, but the Apache 
licence is incompatible because it demands that publicity materials expressly mention 
that combined work contains code of each and every one of the right holders. This does 
not imply that programs with both licences cannot be used simultaneously, or even be 
integrated. It just means that those integrated programs cannot be distributed, since it is 
impossible to comply simultaneously with the redistribution conditions of both. 

The GPL licence is designed to guarantee the freedom of source code at all 
times, since a program published and licensed under its conditions may nev- 
er become proprietary. Moreover, neither that program nor modifications of 
it may be published with a different licence other than the GPL. As already 
mentioned, supporters of the BSD-type licences see a limitation of freedom in 
that clause, whereas followers of the GPL believe that it is a way of ensuring 
that the software will always be free. One way of looking at it would be to con- 
sider that the GPL licence maximises the freedom of users, whereas the BSD 
licence maximises the freedom of developers. However, we should note that 
in the second case we are referring to developers in general and not to authors, 

GNUFDL' PID 00148386 


Free Software 

since many authors consider the GPL licence to be more in their interest, since 
it obliges competitors to publish their modifications (improvements, correc- 
tions, etc.) if they redistribute their software, whereas with a BSD-type licence 
this is not necessarily the case. 

As regards this licence's contrary nature to copyright, it is because its phi- 
losophy (and that of the Free Software Foundation) is that software should 
not have owners (Richard Stallman, "Why software should not have owners", 
1998) [207]. Although it is true that software licensed under the GPL has an 
author, who is the person that allows copyright legislation to apply to it, the 
conditions with which it is published confer such a nature on the software 
that we can consider ownership to correspond to the person in possession of 
the software and not to the person who has created it. 

Of course, this licence also includes disclaimers in order to protect the authors. 
Likewise, in order to protect the good reputation of the original authors, any 
modification of a source file must include a note specifying the date and au- 
thor of the modification. 

The GPL also takes software patents into account, and demands that if the 
source carries patented algorithms (as we have said, something common and 
legal in the US, but currently irregular in Europe), either a licence for use of 
the patent free of charge must be granted, or it cannot be distributed under 
the GPL. 

The latest version of the GPL licence, the second one, was published in 1991 
(although at the time of writing the third one is in an advanced stage of prepa- 
ration). Specifically bearing in mind future versions, the licence recommends 
licensing under the conditions of the second one or any other subsequent one 
published by the Free Software Foundation, which is what many authors do. 
Others, however, including in particular Linus Torvalds (Linux creator), only 
publish their software under the conditions of the second version of the GPL, 
in a bid to distance himself from potential future evolutions of the Free Soft- 
ware Foundation. 

The third version of the GPL ( [115] intends to update it 
to the current software scenario, in respect mainly of aspects such as patents, 
DRM (digital rights management) and other limitations on software freedom. 
For example, the draft currently available at the time of writing (May 2007) 
does not allow a hardware manufacturer to block the use of certain software 
modules if they do not carry a digital signature accrediting a determined au- 
thor. An example of this practice occurs with digital video recorders TiVo, 
which provide the source code to all their software (licenced with GPLv2) at 
the same time as they do not allow modifications of the code to be used on 
the hardware 4 . 

'This case has even suggested the 


use of the word tivoisation for oth- 
er similar cases that have occurred. 

GNUFDL» PID_001 48386 52 Free Software 

Neither does the licence allow the software to be forcibly run on a pre-set 
environment, such as occurs when the use of unsigned kernels is prohibited 
on distributions that consider it appropriate for security reasons. 


There are several points in the GPLv3 licence that have generated a degree of opposition. 
One of the opposing groups is the group of Linux kernel developers (including Linus Tor- 
valds himself). They consider that the requirement to use signed software components 
allows certain security features to be granted that would otherwise be impossible, at the 
same time as their explicit prohibition would extend the licence to the hardware field. 
Plus, the limitation established by the signatures mechanism would only occur with the 
hardware and software platforms that are designed that way, meaning that the software 
could be modified for its use on different hardware. In respect of this point, the FSF is in 
favour of using signature mechanisms that advise against using unsigned components 
for security reasons, but believes that not prohibiting those signatures mechanisms that 
prevent the use of unsigned components, could give rise to scenarios where there would 
be no hardware or software platforms on which to run those software modifications, 
meaning that the liberty of free software would then become totally limited where mod- 
ifying code is concerned. 

The GNU Lesser General Public Licence (GNU LGPL) 

The GNU Lesser General Public Licence, (Free Software Foundation, GNU 
LGPL, version 2.1, February 1999) [119], commonly referred to by its initials 
- LGPL - is another licence of the Free Software Foundation. Designed initial- 
ly for its use with libraries (the L, originally stood for library), it was recently 
modified to be considered the little sister (lesser) of the GPL. 

The LGPL allows free programs to be used with proprietary software. The pro- 
gram itself is redistributed as if it were under the GPL licence, but its integra- 
tion with any other software package is allowed without virtually any restric- 

As we can see, originally this licence was aimed at libraries, in order to promote 
their use and development without encountering the integration problems 
implied by the GPL. However, when it was realised that the pursued objec- 
tive of making free libraries popular was not compensated by the generation 
of free programs, the Free Software Foundation decided to change the library 
to lesser and advised against its use, except in very occasional and particular 
circumstances. Nowadays, there are many programs that are not libraries li- 
censed under the terms of the LGPL. For example, the Mozilla navigator or office suite are also licensed, among others, under the LGPL. 


As is the case with the GPL, the last published version of the LGPL is the second, al- 
though there is already a template of the third version ( 
info-gplv3/2006-July/000008.html) [116]. This new version is shorter than the previous 
one since it refers all its text to the GPLv3 and merely highlights its differences. 

Other strong licences 

Other strong licences that deserve mentioning: 

GNUFDL" PID_001 48386 53 Free Software 

• Sleepycat license ( [59]. 
This is the license under which the company Sleepycat (http:// [60] distributes its programs (of which we could 
mention the well-known Berkeley DB). It enforces certain conditions 
whenever the program or works derived from it are redistributed. In par- 
ticular, it obliges the source code to be offered (including the modifica- 
tions in the case of a derived work) and for the redistribution to impose the 
same conditions on the receiver. Although much shorter than the GNU 
GPL, it is very similar in its main effects. 

• eCos License 2.0 ( [25]. 
This is the license under which eCos ( 
[24], a real-time operating system, is distributed. It is a modification of the 
GNU GPL which does not consider that code linked to the programs it 
protects, is subject to the clauses of the GNU GPL if redistributed. From 
this point of view, its effects are similar to those of the GNU LGPL. 

• Affero General Public License ( [78]. 

It is an interesting modification of the GNU GPL which considers the case 
of programs offering services via the web, or in general, via computer net- 
works. This type of program represents a problem from the point of view of 
strong licences. Since use of the program does not imply having received 
it through a redistribution, even though it is licensed, under the GNU GPL 
for example, someone can modify it and offer a service on the Web us- 
ing it, without redistributing it in any way, and therefore, without being 
obliged, for example, to distribute its source code. The Affero GPL has a 
clause obliging that if the program has a means for providing its source 
code via the web to whoever uses it; this feature may not be disabled. This 
means that if the original author includes this capability in the source 
code, any user can obtain it, and plus that redistribution is subject to the 
conditions of the licence. The Free Software Foundation is considering in- 
cluding similar provisions in version 3 of its GNU GPL. 

• IBM Public License 1.0 ( 
opensource/licenselO.html) [40]. 

It is a licence that allows a binary redistribution of derived works only if 
(among other conditions) a mechanism is contemplated for the person 
receiving the program to receive the source code. The redistribution of 
source code must be made under the same licence. This licence is also 
interesting because it obliges the party redistributing the program with 
modifications, to license automatically and free of charge any patents af- 
fecting such modifications and that are the property of the redistributor 
to the party receiving the program. 

• Mozilla Public License 1.1 (http://www.mozilla.Org/MPL/MPL-l.l.html) 

GNUFDL» PID_001 48386 54 Free Software 

This is an example of a free licence drawn up by a company. It is an evolu- 
tion of the first free licence that Netscape Navigator had, which was very 
important in its day because it was the first time that a well-known com- 
pany decided to distribute a program under its own free licence. 

3.2.4. Distribution under several licences 

Up until now we have assumed that every program is distributed under a sin- 
gle licence which specifies the conditions for use and redistribution. Howev- 
er, an author can distribute works under different licences. In order to under- 
stand this, we should remember that every publication is a new work, and 
that different versions can be distributed with the only difference being in 
their licence. As we will see, most of the time this translates into the fact that 
depending on what the user wants to do with the software he will have to 
observe the terms of one licence or another. 

One of the best known examples of a double licence is the one for the Qt li- 
brary, on which the KDE desktop environment is founded. Trolltech, a com- 
pany based in Norway, distributed Qt with a proprietary licence, although 
it waived payment for programs that didn't use it for profit. For this reason 
and because of its technical characteristics, it was the KDE project's choice in 
the mid-nineties. This gave rise to an intense controversy with the Free Soft- 
ware Foundation because then KDE stopped being completely free software, 
as it depended on a proprietary library. Following an extensive debate (dur- 
ing which GNOME appeared as KDE's free competitor in the desktop environ- 
ment), Trolltech decided to use the double-licence system for its star product: 
the programs under the GPL could use a Qt GPL version, whereas if the in- 
tention was to integrate it with programs that had incompatible licences with 
the GPL (such as proprietary licences), a special licence had to be bought from 
them. This solution satisfied all parties, and nowadays KDE is considered free 

Other well-known examples of dual licences are StarOffice and, or Netscape Communicator and Mozilla. In both cases, the 
first product is proprietary whereas the second is a free version (generally un- 
der the conditions of several free licences). Although originally free projects 
were limited versions of their proprietary siblings, over time they have fol- 
lowed their own path, meaning that nowadays they have a fairly high level 
of independence. 

3.2.5. Program documentation 

The documentation that comes with a program forms an integral part of it, as 
do the comments on source code, as recognised, for example by the Spanish 
Law on Intellectual Property. Given this level of integration, it would seem 

GNUFDL" PID_001 48386 55 Free Software 

logical that the same freedoms should apply to the documentation and that 
it should evolve in the same way as the program: any modification made in a 
program requires a simultaneous and consistent change in its documentation. 

Most of this documentation tends to be coded as unformatted text files, since 
the aim is to make it universally accessible with a minimum tools environ- 
ment, and (in the case of free programs) normally includes a small introduc- 
tion to the program (readme), installation guidelines (install), some history 
on the evolution and future of the program (changelog and todo), authors 
and copyright (authors and copyright or copying), as well as the instruc- 
tions for use. All of these, excluding authors and copyright, must be freely 
modifiable as the program evolves. To authors we just need to add names and 
credits without eliminating anything, and the copyright must only be modi- 
fied if the conditions allow it. 

The instructions for use are normally coded in more complex formats, since 
they tend to be longer and richer documents. Free software demands that this 
documentation may be changed easily, which in turn enforces the use of so- 
called transparent formats, with known specifications and able to be processed 
by free tools, such as, in addition to pure and clean text, the format of the 
Unix manual pages, Texlnfo, LaTeX or DocBook, without prejudice to also be- 
ing able to distribute the result of transforming these source documents into 
more suitable formats for visualisation or printing, such as HTML, PDF or RTF 
(normally more opaque formats). 

However, program documentation is often prepared by third parties who have 
not been involved in the development. Sometimes the documentation is of 
a didactic nature, to facilitate the installation and use of a specific program 
(HO WTO); sometimes it is more extensive documentation that covers several 
programs and their integration, that compares solutions, etc., either in the 
form of a tutorial or reference manual; sometimes it is a mere compilation of 
frequently asked questions and their answers (FAQ). A noteworthy example is 
the Linux documentation project ( [44]. In this category 
we could also include other technical documents, not necessarily about pro- 
grams, whether the instructions for cabling a local network, making a solar 
oven, repairing an engine or selecting a tools supplier. 

These documents are halfway between mere program documentation and 
highly technical and practical articles or books. Without prejudice to the free- 
dom to read, copy, modify and redistribute, the author may wish to give opin- 
ions that he does not want to be distorted, or at least not want any distortion 
to be attributed to him; or he may wish to keep paragraphs, such as acknowl- 
edgements; or make it forcible to modify others, such as the title. Although 
these concerns can also arise with software itself, they have not been expressed 
as vehemently in the world of free software as in the world of free documen- 

GNUFDL» PID_001 48386 56 Free Software 

3.3. Summary 

In this chapter we have looked at the legal aspects that govern or influence 
the world of free software. They form part of intellectual or industrial prop- 
erty legislation conceived, in principle, to stimulate creativity by rewarding 
creators for a specific period. Of the different types, so-called copyright is the 
one that most affects free software, and properly applied it can be used to 
guarantee the existence of free software in the form of free licences. 

We have seen the importance of licences in the world of free software. And 
we have also presented the enormous variety of existing licences, the grounds 
on which they are based, their repercussions, advantages and disadvantages. 
Certainly, we can say that the GPL tries to maximise the freedom of software 
users, whether they receive the free software directly from its author or not, 
whereas BSD-type licences try to maximise the freedom of the modifier or 

Given what we have seen in this chapter, we can deduce that it is very im- 
portant to decide early on what licence a project will have and to be fully 
aware of its advantages and disadvantages, since a later modification tends 
to be extremely difficult, especially if the number of external contributions 
is very large. 

To conclude, we would like to highlight the fact that free software and propri- 
etary software differ solely and exclusively in terms of the licence under which 
the programs are published. In the following chapters, however, we will see 
that this purely legal difference may or may not affect the way software is de- 
veloped, giving rise to a new development model, which can differ from the 
"traditional" development methods used in the software industry to a greater 
or lesser extent, depending on each case. 

GNUFDL» PID 00148386 57 Free Software 

4. Developers and their motivations 

"Being a hacker is lots of fun, but it's a kind of fun that takes a lot of effort. The effort 
takes motivation. Successful athletes get their motivation from a kind of physical delight 
in making their bodies perform, in pushing themselves past their own physical limits. 
Similarly, to be a hacker you have to get a basic thrill from solving problems, sharpening 
your skills and exercising your intelligence." 

Eric Steven Raymond, "How to become a hacker" 

4.1. Introduction 

The partly anonymous and distributed way in which free software is devel- 
oped has meant that for many years the human resources that it relies on 
have been largely unknown. The result of this lack of knowledge has been to 
mythologise, at least to some extent, the world of free software and the life 
of those behind it, based on more or less broad stereotypes about the hacker 
culture and computers. In the last few years, the scientific community has 
made an enormous effort to get to know the people who participate in free 
software projects better, their motivations, academic backgrounds, and other 
potentially relevant aspects. From a purely pragmatic point of view, knowing 
who is involved in this type of projects and why, can be extremely useful when 
it comes to generating free software. Some scientists, mainly economists, psy- 
chologists, and sociologists, have wanted to go further and have seen in this 
community the seed of future virtual communities with their own rules and 
hierarchies, in many cases totally different to those we know in "traditional" 
society. One of the most important mysteries to resolve was to learn what 
motivated software developers to participate in a community of this nature, 
given the fact that financial benefits, at least direct ones, are practically non- 
existent, whereas indirect ones are difficult to quantify. 

4.2. Who are developers? 

This section aims to provide a global overview of the people who spend their 
time and energy participating in free software projects. The data that we show 
stems mostly from scientific research in the last few years, the most significant 
but by no means exclusive including Free/libre and open source software. Survey 
and study, part IV: "Survey of developers", 2002 [126], and "Who is doing it? 
Knowing more about libre software developers", 2001 [197]. 

Software developers are normally young people. The average age is around 
twenty seven. The variation in age is significant, since the dominant group 
is in the twenty one to twenty four age bracket, and the most frequently ap- 
pearing value is twenty three. It is interesting to note how the age of joining 
the free software movement peaks between eighteen and twenty five and is 
particularly pronounced between twenty one and twenty three, which would 

GNUFDL» PID_001 48386 58 Free Software 

coincide with university age. This evidence stands in contrast to the claim that 
free software is mostly a teenage thing, although there is an obvious involve- 
ment of teenagers (about 20% of developers are under twenty). For certain, 
what we can see is that most developers (60%) are in their twenties, with the 
under-twenties and over-thirties equally sharing the remaining 40%. 

From the age of joining we can deduce that there is an enormous universi- 
ty influence on free software. This is not surprising, given that as we have 
seen in the chapter on history, before free software was even known by that 
name it was closely connected to higher education. Even today, student user 
groups and universities continue to drive the use and expansion of free soft- 
ware. Therefore, it is not surprising that more than 70% of developers have a 
university education. This data is even more significant if we bear in mind that 
the remaining 30% are not at university yet because they are still in school. 
Even so, they are also involved and are no less appreciated than developers 
who have never had access to higher education, but are IT enthusiasts. 

The free software developer is normally male. The figures juggled by different 
surveys on the presence of women in the community vary between 1% and 
3%, competing with their own margin of error. At the same time, a majority 
(60%) claims to have a partner, while the number of developers with children 
is just 16%. Given the age brackets of free software developers, this data coin- 
cides fairly accurately with a random sample, meaning that it may be consid- 
ered "normal". The myth of the lonely developer whose enthusiasm for IT is 
the only thing in his life is shown to be, as we can see, an exception rather 
than the rule. 

4.3. What do developers do? 

Professionally speaking, free software developers describe themselves as soft- 
ware engineers (33%), students (21%), programmers (11%), consultants 
(10%), university professors (7%), etc. On the opposite end of the scale, we 
find that they tend not to form part of sales or marketing departments (about 
1%). It is interesting to note how many of them define themselves as software 
engineers rather than programmers, almost three times as many, bearing in 
mind, as we will see in the chapter on software engineering, that the appli- 
cation of classical software engineering techniques (and even some modern 
ones) is not really entrenched in the world of free software. 

The university connection, which has already been proven, rears its head 
again in this section. About one in three developers is a student or university 
professor, which goes to show the tremendous collaboration between people 
mainly from the software industry (the remaining two thirds) and the aca- 
demic sphere. 

GNUFDL» PID_001 48386 59 Free Software 

At the same time, it has been possible to identify a large scope of mixed disci- 
plines: one in five developers comes from a field that is not IT. This, combined 
with the fact that there is also a similar number of non-university developers, 
reflects a wealth of interests, origins, and certainly, composition of develop- 
ment teams. It is very difficult to find a modern industry, if there is one, where 
the degree of heterogeneity is as large as the one we can see in free software. 

In addition to the approximately 20% of students, 64% of developers are most- 
ly paid employees, whereas the proportion of self-employed developers is 14%. 
Finally, just 3% claims to be unemployed, a significant fact since the survey 
was conducted after the dotcom crisis began. 


The fact that free software business model, unlike with proprietary software, cannot be 
achieved through the sale of licences has always propitiated heated debates as to how 
programmers should earn their living. In the surveys that we refer to in this chapter, 50% 
of developers claimed to have obtained direct or indirect financial compensation for their 
involvement in free software. However, many others aren't so sure. Richard Stallman, 
founder of the GNU project, when asked what a free software developer should do in 
order to make money, tends to reply that he can work as a waiter. 

4.4. Geographical distribution 

Obtaining developers' geographical data is an issue that needs to be ap- 
proached in a more scientific manner. The problem with the research shown 
in this chapter is that because it is based on Internet surveys, open to anyone 
wishing to participate, participation has depended to a great extent on the 
sites it was posted, and the way in which it was announced. To be accurate, we 
should note that the surveys did not aim to be representative in this regard, 
but rather to obtain the answers and/or opinions of the largest possible num- 
ber of free software developers. 

However, we could venture to make a few statements on this issue, know- 
ing that this data is not as reliable as previous data, and that therefore, the 
margin of error is much greater. What seems to be a fact is that most free 
software developers come from industrialised countries, and that the pres- 
ence of developers from so-called Third World countries is rare. Consequently, 
it shouldn't be surprising that the map of developers of the Debian project 
( [187], for example, matches the 
photographs of the Earth at night: where there is light - read "where there is 
an industrialised civilisation" - that is where free software developers tend to 
concentrate. This, which may seem logical at first sight, stands in contrast to 
the potential opportunities that free software can offer Third World countries. 

We can find a clear example in the following table, which contains the most 
common countries of origin of the Debian project developers in the last four 
years. There is a noticeable trend towards decentralisation of the project, ev- 
ident from the fact that the growth in the number of developers from the 
US, the country which most contributes, is lower than the average. And the 

GNUFDL* PID 00148386 


Free Software 

fact is that, in general, countries have managed to double their numbers of 
volunteers from 1999 to 2003, and France, which has managed to multiply 
its presence by five, is the clearest example in this regard. Considering that 
Debian took its first steps on the American continent (in the US and in Cana- 
da in particular), we can see that in the last few years the project has become 
Europeanised. We assume that the following step will be the much sought-after 
globalisation, with the incorporation of South American, African and Asian 
countries (with the exception of Korea and Japan, which are already well rep- 
resented), although the data in our possession (two collaborators from Egypt, 
China or India, and one in Mexico, Turkey or Colombia in June 2003) are not 
very promising in this sense. 

In the world of free software (and not just in the case of Debian), there is an 
extensive debate over the supremacy of Europe or the United States. Almost 
all studies have shown that the presence of European developers is slightly 
higher than the North American one, an effect that is mitigated by the fact 
that Europe's population is greater than that of the US. Therefore, we are deal- 
ing with a war of numbers, since the number of developers per capita is higher 
among the North Americans. If we take into account the number of people 
with Internet access instead of the total population, then Europe comes out 
on top again. 

In terms of countries, the areas with the highest levels of implantation (in 
numbers of developers divided by the population) are Northern Europe (Fin- 
land, Sweden, Norway, Denmark and Iceland) and Central Europe (Benelux, 
Germany and the Czech Republic), followed by Australia, Canada, New 
Zealand and the US. Despite its importance in absolute terms (due to the large 
populations of France, Italy and Spain), the Mediterranean zone however, is 
below average. 

Table 1. Countries with the largest number of Debian developers 





























































GNUFDL* PID 00148386 


Free Software 



















4.5. Dedication 

The number of hours that free software developers spend on developing free 
software is one of the big unknowns. We should also point out that this is one 
of the main differences with company-generated software, where the members 
of the team and the time spent by each team member on a development are 
well known. The time that developers dedicate to free software can be taken 
as an indirect measure of their level of professionalisation. Before showing 
the data currently available, it is important to note that it has been obtained 
from estimates given by developers in various surveys, so that in addition to 
the inherent inaccuracies of this type of data gathering, we need to consider 
a margin of error associated to how each developer interprets development 
time. Hence, it is certain that many developers will not include the time spent 
reading e-mails (or perhaps they will) and only specify the time they spend 
programming and debugging. Therefore, all the figures we show next need to 
be considered with due reserve. 

The research conducted to date shows that each software developer spends 
eleven hours a week on average ("Motivation of software developers in open 
source projects: an internet -based survey of contributors to the Linux kernel", 
2003) [143]. However, this figure can be deceptive, since there is an enormous 
variation in the time spent by software developers. In the study Free/libre and 
open source software. Survey and study, part IV: "Survey of developers", 2002 
[126], 22.5% of those surveyed said that their contribution was less than two 
hours per week, and this figure increased to 26.5% for those spending two to 
five hours per week; between six and ten hours was the time spent by 21.0%, 
while 14.1% spent between eleven and twenty hours per week; 9.2% claimed 
that the time they spent developing free software was between twenty and 
forty hours per week and 7.1%, over forty hours per week. 

Table 2. Dedication in hours per week 

Hours per week 


Less than 2 hours 


Between 2 and 5 hours 


Between 5 and 10 hours 

21 .0% 

Between 10 and 20 hours 


GNUFDL* PID 00148386 62 Free Software 

Hours per week 


Between 20 and 40 hours 


More than 40 hours 



In addition to showing the level of professionalisation of free software development 
teams, the time spent in hours is a relevant parameter when it comes to making cost 
estimates and comparisons with proprietary development models in the industry. With 
free software, for now, we just have the end products (new software deliveries, synchro- 
nisation of new code in versions systems...) which do not allow us to know how much 
time the developer has spent on achieving them. 

An analysis of these figures tell us that about 80% of developers perform these 
tasks in their free time, whereas only one in five could consider that they 
spend as much time on this activity as a professional. Later, in the chapter on 
software engineering, we will see how this data matches developers' contribu- 
tions, since they both appear to follow the Pareto law (vid. section 7.6). 

4.6. Motivations 

There has been much speculation as to the motivations that developers have 
to develop free software, especially when it is done in free time (which, as we 
have seen, corresponds to about 80% of developers). As in previous sections, 
we only have the survey data, which is why it is important to realise that the 
answers have been given by the developers themselves, meaning that they 
may be more or less coherent with reality The percentages shown next exceed 
the 100% mark because there was an option for participants to select several 

In any case, it appears from their answers that most want to learn and to de- 
velop new skills (approximately 80%) and that many do so in order to share 
knowledge and skills (50%) or to participate in a new form of cooperation 
(about a third). The first data is not surprising, given that a professional with 
more knowledge will be in greater demand than one with less. However, it 
is not quite so easy to explain the second data, and it would even seem to 
contradict Nikolai Bezroukov's opinion in "A second look at the cathedral and 
the bazaar" (December, 1998) [91] that the leaders of free software projects are 
careful not to share all the information in their possession in order to perpet- 
uate their power. Meanwhile, the third most frequent choice is undoubtedly, a 
true reflection of developers' enthusiasm about the way free software is created 
in general; it is difficult to find an industry in which a group of lightly organ- 
ised volunteers can -technologically speaking- stand up to the sector's giants. 

Although the "classical" theory for explaining why free software developers 
spend time contributing to this type of projects is reputation and indirect fi- 
nancial benefits in the medium and long term, it would appear that developers 
themselves disagree with these claims. Just 5% of those surveyed replied that 

GNUFDL* PID 00148386 


Free Software 

they develop free software in order to make money, whereas the number who 
did so in order to establish a reputation rose to 9%, far from the answers given 
in the preceding paragraph. In any case, it seems that researching developers' 
motivations to become part of the free software community is a fundamental 
task, which sociologists and psychologists will have to face in the near future. 

4.7. Leadership 

Reputation and leadership are two of the characteristics used to explain the 
success of free software, and especially, the bazaar model, as we will see from 
the chapter on software engineering. As we have seen in another chapter, on 
software licences, there are certain differences between free software licences 
and its equivalents in the documentation field. These differences stem from 
the way of retaining authorship and authors' more accentuated opinion in 
text than in programs. 

In Freefiibre and open source software. Survey and study, part IV: "Survey of de- 
velopers" (2002) [126] a question was included that asked developers to point 
out what people from a list were known to them, not necessarily personally. 
The results, set out in table 3, show that these people can be divided into three 
clearly distinct groups: 

Table 3. Level of awareness of important developers 


Known for 

Linus Torvalds 


Richard Stallman 


Miguel de Icaza 


Eric Raymond 


Bruce Perens 


Jamie Zawinski 


Mathias Ettrich 


Jorg Schilling 


Marco Pesenti Gritti 


Bryan Andrews 


Guenter Bartsch 


Arpad Gereoffy 


Martin Hoffstede 


Angelo Roulini 


Sal Valliger 

1 .2% 

GNUFDL» PID 00148386 64 Free Software 

• A first group of people with clear philosophical and historical connota- 
tions within the world of free software (although, as we know, they may 
also have notable technical skills): 

1) Linus Torvalds. Creator of the Linux kernel, the most used free operating 
system, and co-author of Just for fun: the story of an accidental revolutionary 

2) Richard Stallman. Ideologist and founder of the Free Software Foundation 
and developer on various GNU projects. Author of several important es- 
says on free software ("Why free software is better than open source", 1998 
[206], "Copyleft: pragmatic idealism", 1998 [205], "The GNU Project" [208] 
and "The GNU Manifesto", 1985 [117]). 

3) Miguel de Icaza. Co-founder of the GNOME project and Ximian Inc., and 
developer in GNOME and MONO. 

4) Eric Raymond. Promoter of the Open Source Initiative, author of "The 
cathedral and the bazaar" [192] and main developer of fetchmail. 

5) Bruce Perens. Former leader of the Debian project, promoter (converted) 
of the Open Source Initiative and developer of the e-fence tool. 

6) Jamie Zawinsky. Ex developer of Mozilla and famous for a letter of 1999 
in which he left the Mozilla project arguing that the model they were 
following would never bear fruit ("Resignation and postmortem", 1999) 


7) Mathias Ettrich. Founder of KDE and developer of LyX and others. 

• A second group consisting of developers. This survey took the names of 
the leading developers of the six most popular projects according to the 
FreshMeat free software download site. We can see that (with the excep- 
tion of Linus Torvalds, for obvious reasons, and Jorg Schilling) the level of 
awareness of these developers is small: 

1) Jorg Schilling, creator of cdrecord, among other applications. 

2) Marco Pesenti Gritti, main developer of Galeon. 

3) Bryan Andrews, developer of Apache Toolbox. 

4) Guenther Bartsch, creator of Xine. 

GNUFDL» PID_001 48386 65 Free Software 

5) Arpad Gereoffy, developer of MPEGPlayer. 

• A third group consisting of the names of the three last "people" in the 
table. These names were invented by the survey team in order to check 
the margin of error. 

We can draw two conclusions from the results: the first is that we can consider 
the margin of error to be small (less than 3%), and the second is that most 
developers of the most popular free software applications are as well-known 
as people who do not exist. This data should make those who allege that one 
of the main reasons for developing free software is fame-seeking think twice. 

4.8. Summary and conclusions 

This chapter has attempted to shed some light on the largely unknown issue 
of the people who dedicate time to free software. In general terms, we can say 
that a free software developer is a young male with a university qualification 
(or on the way to getting one). The relationship between the world of free 
software and universities (students and professors) is very close, although the 
developer who has nothing to do with the academic sphere still predominates. 

In terms of hours of dedication, we have shown that there is an enormous 
disparity, similar to what is postulated in the Pareto law. Developers' moti- 
vations, in their own opinion, are far from being monetary and egocentric, 
as economists and psychologists tend to suggest, and are mostly to do with 
sharing and learning. Finally, we have shown a table of the most significant 
personalities in the world of free software (including others who were not so 
well-known, as we have seen) and shown that reputation in the enormous free 
software community tends to depend on more than just coding a successful 
free software application. 

GNUFDL» PID 00148386 66 Free Software 

5. Economy 

"Res publica non dominetur." 

"Public things have no owner." (free translation) 

Appeared in an IBM advert about Linux (2003) 

This chapter looks at some economic aspects related to free software. We will 
start by showing how free software projects are financed (when they are in- 
deed financed, since in many cases they rely solely on efforts and resources 
contributed voluntarily). Next, we will look at the main business models put 
into practice by companies directly associated to free software. The chapter 
ends with a small study of the relationship between free software and monop- 
olies in the software industry. 

5.1. Funding free software projects 

Free software is developed in many different ways and using mechanisms to 
obtain funds that vary enormously from case to case. Every free project has 
its own way of financing itself, from the one consisting totally of volunteer 
developers and using only altruistically ceded funds, to the one carried out by 
a company that invoices 100% of its costs to an organisation interested in the 
corresponding development. 

In this section, we will focus on the projects where there is external funding 
and not all the work is voluntary. In these cases, there is normally some form of 
cash inflow, from an external source to the project, responsible for providing 
funds for its development. This way, the free software that is built may be 
considered, to some extent, to be the product of this external funding. This 
is why it is common for that external source to decide (at least in part) how 
the funds are spent and on what. 

In some way, this external funding for free projects can be considered a kind of 
sponsorship, although this sponsorship has no reason for being disinterested 
(and usually it is not). In the following sections we discuss the most common 
types of external funding. However, while learning about them, we should re- 
member that these are just some of the ways that free software projects obtain 
resources. But there are others, and of these the most important one (as we 
have seen in chapter 4) is the work of many volunteer developers. 

5.1.1. Public funding 

A very special type of financing for free projects is public funding. The funding 
body may be directly a government (local, regional, national or even supra- 
national) or a public institution (for example, a foundation). In these cases, 

GNUFDL» PID_001 48386 67 Free Software 

the funding tends to be similar as for research and development projects, and 
in fact it is common for the funding to come from public bodies that promote 
R+D. Normally, the funding body will not seek to recover the investment (or at 
least not directly), although it tends to have clear objectives (such as promot- 
ing the creation of an industrial or research-based fabric, promoting a certain 
technology or a certain type of application, etc.). 

In most of these cases, there is no explicit financing of products or services 
related to free software, but rather this tends to be the sub-product of a con- 
tract with other more general objectives. For example, as part of its research 
programs, the European Commission funds projects aimed at improving Eu- 
ropean competitiveness in certain fields. Some of these projects have as part 
of their objectives to use, improve and create free software within the scope 
of the research (as a research tool or a product derived from it). 

The motivations for this type of financing are very varied, but we can distin- 
guish the following: 

1) Scientific. This is the most frequent one in the case of publicly funded 
research projects. Although its objective is not to produce software but 
rather to investigate a specific field (whether IT-related or not), it is likely 
to require programs to be developed as tools for achieving the project's 
objectives. Usually the project is not interested in commercialising these 
tools, or may even be actively interested in other groups using and im- 
proving them. In such cases, it is fairly common to distribute them as free 
software. In this way, the group conducting the research has partly ded- 
icated funds to producing this software, so we can say that it has been 
developed with public funding. 

2) Promoting standards. Having a reference implementation is one of the 
best ways of promoting a standard. In many cases this involves having 
programs that form part of said implementation (or if the standard refers 
to the software field, to be the implementation themselves). For the refer- 
ence implementation to be useful in promoting the standard, it needs to 
be available, at least in order to check interoperativity for all those wishing 
to develop products that subscribe to that standard. And in many cases 
it is also advisable for manufacturers to be able to adapt the reference im- 
plementation directly in order to use it with their products if they wish. 
This is how, for example, the Internet protocols were developed, which 
have now become a universal norm. In such cases, releasing reference im- 
plementations as free software can contribute enormously towards that 
promotion. Once again, free software here is a sub-product, in the case of 
promoting a standard. And normally the party responsible for this promo- 
tion is a public body (although sometimes it may be a private consortium). 

3) Social. Free software is a very interesting tool for creating the basic in- 
frastructure for the information society. Organisations interested in using 

GNUFDL» PID_001 48386 68 Free Software 

free software to promote universal access to the information society can 
finance projects related to it (normally with projects for developing new 
applications or adapting already existing ones). 


An example of public financing for a primarily social objective is the case of gnuLinEx, 
promed by the Extremadura Regional Government (Extremadura, Spain) in order to pro- 
mote the information society fundamentally in terms of computer literacy. The Regional 
Government has financed the development of a distribution based on Debian in order 
to achieve this objective. Another similar case is the German government funding of 
GnuPG developments, aimed at making them easier to use for inexperienced users, with 
the idea of promoting the use of secure mail by its citizens. 

The development of GNAT 

A notorious case of public financing for a free software development is the case of the 
GNAT compiler. GNAT, the Ada compiler, was financed by the Ada 9X project of the 
US Department of Defence, with the idea of having a compiler of the new version of 
the Ada programming language (which would later become Ada 95), which it was trying 
to promote at that time. One of the causes identified in relation to software companies 
adopting Ada's first version (Ada 83) was the late availability of a language compiler and 
its high cost when it was finally released. Therefore, they tried to prevent the same thing 
from happening with Ada 95, ensuring that the compiler was ready almost simultane- 
ously with the release of the language's new standard. 

To do so, Ada 9X contracted a project with a team from the University of New York (NYU), 
for an approximate value of one million USD, to carry out a "concept implementation" 
of the Ada 95 compiler. Using these funds, and taking advantage of the existence of GCC 
(GNU's C compiler, of which most of the backend was used), the NYU team effectively 
built the first Ada 95 compiler, which it released under the GNU GPL. The compiler was 
so successful that when the project was finished some of its creators founded a company 
(Ada Core Technologies), which since then has become the market leader in compilers 
and help tools for building programs in Ada. 

In this project it is worthy to note the combination of research elements (in fact, this 
project advanced knowledge on the building of front ends and run time systems for 
Ada-type language compilers) and promotion of standards (which was the funding body's 
clearest objective). 

5.1.2. Private not-for-profit funding 

This type of funding has many similar characteristics to the previous type, 
which is normally conducted by foundations or non-governmental organisa- 
tions. Direct motivation in these cases tends to be to produce free software 
for use in a sphere that the funding body considers particularly relevant, but 
we may also find the indirect motivation of contributing to problem-solving 
(for example, a foundation that promotes research into a disease may finance 
the construction of a statistics program that helps to analyse the experimental 
groups used as part of the research into that disease). 

In general, both the motives and the mechanisms for this type of funding are 
very similar to those of public funding, although naturally they are always in 
the context of the funding body's objectives. 

GNUFDL» PID_001 48386 69 Free Software 


Probably, the archetypal case of a foundation that promotes the development of free 
software is the Free Software Foundation (FSF). Since the mid-1980s this foundation is 
dedicated to promoting the GNU project and to encouraging the development of free 
software in general. 

Another interesting case, although in a rather separate field, is the Open Bioinformatics 
Foundation. The objectives of this foundation include promoting the development of 
basic computer programs for research in any of the branches of bioinformatics. An in 
general, it promotes this by financing and contributing to the construction of free pro- 

5.1.3. Financing by someone requiring improvements 

Another type of financing the development of free software, which is not quite 
so altruistic, takes place when someone needs to make improvements to a free 
product. For example, for internal use, a company may need a certain program 
to have a particular functionality or to correct a few bugs. In these cases, it is 
common for the company in question to contract the required development. 
This development is often free software (whether because the licence of the 
modified program imposes it, or because the company decides it). 

The case of Corel and Wine 

Towards the end of the 1990s, Corel decided to port its products to GNU/Linux. During 
this process it discovered that a free program designed to facilitate the execution of bina- 
ries for Windows in Linux environments could help to make considerable development 
savings. But in order to do so, it had to be improved, fundamentally by adding the em- 
ulation of some Windows functionality that the Corel programs used. 

For this, Corel contracted Macadamian, which contributed its improvements to the Wine 
project. This way, both Corel and Wine benefited. 

5.1.4. Funding with related benefits 

With this type of financing, the funding body aims to obtain benefits from 
products related to the program whose development it funds. Normally, in 
these cases the benefits obtained by the funding body are not exclusive, since 
others can also enter the market for selling the related products, but either 
the market share it captures is sufficient for it not to worry too much about 
sharing the pie with others, or it has a clear competitive advantage. 

Some examples of products related to a particular software are as follows: 

• Books. The company in question sells manuals, user guides, course mate- 
rials, etc. related to the free program that it helps to finance. Of course, 
other companies can also sell related books, but normally financing the 
project will give the company early access to key developers before the 
competition, or simply provide a good image towards the user community 
of the program in question. 

• Hardware. If a company funds the development of free systems for a cer- 
tain type of hardware, it can more easily dedicate itself to selling that type 

GNUFDL" PID_001 48386 70 Free Software 

of hardware. Once again, since the software developed is free, competitors 
selling the same type of devices may appear, that use the same develop- 
ments without having collaborated in the funding. But even so, the com- 
pany in question has several advantages over its competitors, and one of 
them may be that its position as a source of funding for the project allows 
it to exert influence so that priority is given to the developments in which 
it is most interested. 

• CD with programs. Probably, the best known model of this type is the 
one of companies financing certain developments that they then apply to 
their software distribution. For example, having a good desktop environ- 
ment can help a lot to sell a CD with a certain distribution of GNU/Linux, 
and therefore, financing its development could be a good business for the 
party selling the CDs. 

We need to bear in mind that under this heading the financing in question 
has to be made with a profit motivation, and therefore the funding body has 
to obtain a potential benefit from the financing. In reality, however, it is com- 
mon for there to be a combination of profit motive and altruism when a com- 
pany provides funds for a free program to be made from which it expects to 
benefit indirectly. 


A well-known case of funds contributed to a project, albeit fairly indirectly, is the help 
that the O'Reilly publishing house gives to the development of Perl. Naturally, it is no 
coincidence that O'Reilly is also one of the main publishers of subjects related to Perl. 
In any case, it is obvious that O'Reilly does not have exclusivity over the publication of 
books of this kind, and that other publishing houses compete in this market segment, 
with varying degrees of success. 

VA Software (originally VA Research and later VA Linux) has collaborated actively in de- 
veloping the Linux kernel. Through this, it has achieved, among others, ensured conti- 
nuity, which was particularly critical for it in relation to its customers when its main 
business was selling equipment with a GNU/Linux pre-installation. 

Red Hat has financed the development of many GNOME components, essentially obtain- 
ing a desktop environment for its distribution, which has contributed to increasing its 
sales. As in previous cases, other manufacturers of distributions have benefited from this 
development, although many of them have not collaborated with the GNOME project 
to the same extent as Red Hat (and quite a few have not collaborated at all). Despite this 
fact, Red Hat benefits from its contribution to GNOME. 

5.1.5. Financing as an internal investment 

There are companies that develop free software directly as part of their busi- 
ness model. For example, a company may decide to start a new free project in 
a field where it believes that there are business opportunities, with the idea of 
subsequently obtaining a return on that investment. This model could be con- 
sidered a variation of the previous one (indirect financing), and the "related 
benefits" would be the advantages that the company obtains from producing 
the free program. But since in this case it is the free product itself which is ex- 
pected to produce the benefits, it seems appropriate to give it its own heading. 

GNUFDL» PID_001 48386 71 Free Software 

This type of financing gives rise to various business models. When we anal- 
yse them (in section 5.2) we will also explain the advantages that a company 
normally obtains from this type of investment in a project and what meth- 
ods tend to be used in order to make it profitable. But in any case, we should 
mention that sometimes the software in question may be developed simply 
in order to satisfy the company's own needs, and that only later the company 
may decide to release it, and perhaps, to open a business line based on it. 


Digital Creations (now Zope Corporation) is one of the most well-known cases of a com- 
pany dedicated to developing free software with the expectation of making a return on 
its investment. The free project that Zope invests most heavily in is an applications server 
that is enjoying a certain amount of success. Its history with free software started when 
the then Digital Creations was looking for venture capital to develop its proprietary ap- 
plications server, around 1998. One of the groups most interested in investing in them 
(Opticality Ventures) established as a condition that the resulting product must be free, 
because otherwise they did not see how they could obtain a significant market share. 
Digital Creations agreed to this approach and a few months later announced the first 
version of Zope. Nowadays, Zope Corporation is specialised in consulting, training and 
support for content management systems based on Zope, and other products of which 
Zope is unquestionably the cornerstone. 

Ximian (formerly Helix Code) is a well-known case of free applications development in 
a business environment. Closely linked since its origins to the GNOME project, Ximi- 
an has produced software systems such as Evolution (a personal information manager 
which includes a relatively similar functionality to Microsoft Outlook), Red Carpet (an 
easy-to-use system for managing packages on an operating system) and MONO (an im- 
plementation of a large part of NET). The company was founded in October 1999 and 
attracted many developers from GNOME, who became members of its development team 
(while continuing in many cases to collaborate with the GNOME project). Ximian po- 
sitioned itself as an engineering company expert in GNOME adaptations, in building 
applications based on GNOME, and in general, in providing development services based 
on free software, especially tools related to the desktop environment. In August 2003, 
Ximian was bought by Novell. 

Cisco Enterprise Print System (CEPS) ( [1 7] is a printing man- 
agement system for organisations that use very many printers. It was developed internal- 
ly in Cisco to satisfy its own needs and freed in 2000 under the GNU GPL. It is difficult to 
know for sure Cisco's reasons for doing this, but they could be related to finding external 
contributions (error reports, new controllers, patches, etc.). In any case, what is obvious 
is that since Cisco had no plans to commercialise the product and its potential market 
was not very clear, it had very little to lose with this decision. 

5.1.6. Other financing modes 

There are other financing modes that are difficult to classify under the previ- 
ous headings. As an example, we could mention the following: 

• Use of the market for putting developers and clients in contact. The idea 
that sustains this mode of financing is that, especially for minor develop- 
ments, it is difficult for a client wanting a specific development to come 
into contact with a developer capable of doing it in an efficient manner. 
In order to improve this situation, free software development markets are 
established where developers can advertise their skills and clients, the de- 
velopments that they need. A developer and a client reach an agreement; 
we have a similar situation to the one already described as "funding by a 
party requiring improvements " (section 5.1.3). 

GNUFDL" PID_001 48386 72 Free Software 


SourceXchange is an example of a market that put developers in contact with potential 
clients. To advertise a project, a client would present an RFP (request for proposal) spec- 
ifying the development required and the resources it was prepared to provide for that 
development. These RFPs were published on the site. When a developer read one that 
interested him, he would make an offer for it. If a developer and a client agreed on the 
terms of the development, a project would begin. Normally, every project was supervised 
by a peer reviewer, a reviewer responsible for ensuring that the developer complied with 
the specifications and that indeed the specifications made sense, and advising on how 
to carry through the project, etc. SourceXchange (owned by the company CollabNet) 
took charge of providing the site, guaranteeing reviewers' capabilities, ensuring payment 
in the case of completed projects and offering monitoring tools (services for which it 
invoiced the client). The first project mediated through SourceXchange was completed 
in March 2000, but just over a year later, in April 2001, the site closed down. 

• Project financing through the sale of bonds. The idea behind this type of 
financing is similar to that of the ordinary bonds market approached by 
companies, but targeted at developing free software. It has several varia- 
tions, but one of the best known operates as follows. When a developer 
(an individual or a company) has an idea for a new program, or improve- 
ment for an existing program, he writes it up as a specification, with a 
cost estimate for its development and issues bonds for its construction. 
The value of these bonds is only executed if the project is finally complet- 
ed. When the developer has sold enough bonds, development begins, fi- 
nanced with loans based on them. When the development is completed, 
and an independent third party certifies that indeed what has been done 
complies with the specifications, the developer "executes" the bonds, set- 
tles the debts, and what is left over is the profit made from the develop- 

Who would be interested in buying these bonds? Obviously, users who 
would want the new program or improvement to an existing program to 
be made. To some extent, this bonds system allows interested parties to 
establish developers' priorities (at least in part), through the acquisition of 
bonds. This also means that development costs do not have to be assumed 
by just one company, but rather can be shared between several (including 
individuals), who additionally only have to pay if the project concludes 
successfully in the end. A similar mechanism to this is proposed in much 
more detail in "The Wall Street performer protocol. Using software com- 
pletion bonds to fund open source software development", by Chris Rasch 
(2001) [191]. 


The bonds system described is based on the street performer protocol ("The street performer 
protocol", in: Third USENIX Workshop on Electronic Commerce Proceedings, 1998 [152], and 
"The street performer protocol and digital copyrights", 1999 [153]), a mechanism based 
on e-commerce designed to facilitate private funding of free creations. In short, whoev- 
er is interested in a particular job would formally promise to pay a certain amount if 
the work is done and published as free. Its objective is to find a new way of financing 
relatively small jobs that are made available to everyone, but may be extended in many 
ways (the bonds for the construction of free software being one of them). We can see 
a small case of putting a derivation of this protocol into practice, the rational street per- 
former protocol (Paul Harrison, 2002, [137]) where 
circle/funding_results.htmlit is applied to obtaining funds to finance part of The Circle, 
a free software project. 

GNUFDL» PID_001 48386 73 Free Software 

• Developer cooperatives. In this case, free software developers, instead of 
working individually or for a company, join some form of association (nor- 
mally similar to a cooperative). In all other aspects, it functions the same 
way as a company, with an overtone of its ethical commitment to free 
software, which may form part of its company statutes (although an ordi- 
nary company can do this too). In this type of organisation, we may see 
a variety of combinations of voluntary and paid work. An example is Free 

• Donations system. This involves enabling a mechanism for paying the 
author of a particular software, through the web page that accommodates 
the project. This way, users interested in the project continuing to release 
new versions can support it financially by making voluntary donations in 
the way of funding for the developer. 

5.2. Business models based on free software 

In addition to the project funding mechanisms that we have already talked 
about, another aspect related to the economy which deserves mentioning is 
business models. In speaking about financing mechanisms, we have already 
mentioned a few in passing. Here, in this section, we will describe them in a 
more methodical fashion. 

In general, we can say that many business models are being explored around 
free software, some more classical and others more innovative. We need to 
take into account that it is not easy to use those based on the sale of licences, 
the most common found models in software industry, since in the world of 
free software this financing mechanism is very difficult to exploit. However, 
we can use those based on services to third parties, with the advantage that it 
is possible to offer complete support for a program without necessarily being 
its producer. 

Sale of free software at so much per copy 

In the world of free software it is difficult to charge for licences for use, but not impossible. 
In general, there is nothing in the free software definitions to prevent a company from 
creating a product and only distributing it to anyone who pays a certain amount. For 
example, a particular producer could decide to distribute its product with a free licence, 
but only to whoever pays 1,000 euros per copy (like in the classical world of proprietary 

However, although theoretically this is possible, in practice it is difficult for this to occur. 
Because once the producer has sold the first copy, whoever receives it may be motivated to 
try and recover his or her investment by selling more copies at a lower price (something 
which cannot be prohibited by the program's licence if it is free). In the previous example, 
one could try selling ten copies at 100 euros each, meaning that additionally the product 
would work out free of charge (also, this would make it very difficult for the original 
producer to sell another copy at 1,000 euros, since the product could be legally obtained 
at a tenth of the cost). It is easy to see how this process would continue in waterfall 
until copies were sold at a price close to the copying marginal cost, which with current 
technologies is practically zero. 

GNUFDL» PID_001 48386 74 Free Software 

Even so, and bearing in mind that the mechanism described will mean that normally a 
producer cannot put a price (particularly a high price) on the mere fact of the program's 
redistribution, there are business models that implicitly do just that. One example is the 
case of GNU/Linux distributions, which are sold at a much lower price in comparison 
with proprietary competitors, but above (and normally far above) the cost of the copy 
(even when it can be freely downloaded from the Internet). Of course, in these cases other 
factors come into play, such as the brand image or convenience for the consumer. But 
this is not the only case. Therefore, rather than saying that free software "cannot be sold 
at so much per copy", we should bear in mind that it is more difficult to do so, and that 
probably it will generate less profit, but that there can be models based precisely on that. 

Given these limitations (and these advantages), for several years now varia- 
tions on the usual business models in the software industry are being tried 
out, at the same time as other more innovative solutions are sought for ex- 
ploiting the possibilities offered by free software. No doubt, in the next few 
years we will see even more experimentation in this field, and will also have 
more information on what models can work and under what circumstances. 

In this section we offer a panorama of the business models that we most fre- 
quently encounter today, divided into groups with the intention of showing 
the reader what they share in common and what distinguishes them, focusing 
on those based on the development and services around a free software prod- 
uct. Revenue, in this case, comes directly from the development activities and 
services for the product, but does not necessarily imply new product develop- 
ment. When this development does occur, these models have the financing 
of free software products as a subproduct, meaning that they are particularly 
interesting models with a potentially large impact on the world of free soft- 
ware in general. 

In any case, and although here we offer a relatively clear classification, we 
must not forget that almost all companies in reality use combinations of the 
models that we describe, and with other more traditional ones. 

5.2.1. Better knowledge 

The company that follows this business model tries to make profits on its 
knowledge of a free product (or set of products). Its revenue will come from 
clients to which it will sell services related to that knowledge: development 
based on the product, modification, adaptation, installation and integration 
with other products. The company's competitive advantage will be closely 
related to its better knowledge of the product: therefore, the company will be 
particularly well positioned if it is the producer or an active participant in the 
project producing the software product. 

This is one of the reasons why companies that use this model tend to be active 
participants in the projects related to the software for which they try to sell 
services: it is a very efficient way of obtaining knowledge about it, and more 
importantly, for that knowledge to be recognised. Certainly, being able to tell a 

GNUFDL» PID_001 48386 75 Free Software 

client that the company's employees include various developers on the project 
that produces the software, which, for example, needs to be changed, tends 
to provide a good guarantee. 

Relationship with development projects 

Therefore, companies of this type are very interested in transmitting an image of having 
good knowledge of certain free products. An interesting outcome of this is that support 
for free software projects (for example, by participating actively in them, or allowing 
employees to do so in the course of the working day) is not therefore, something purely 
philanthropic. On the contrary, it may be one of the company's most profitable assets, 
since clients will value it very positively as a clear sign that the company is knowledgeable 
about the product in question. Plus, this way it will be able to follow the development 
closely, trying to make sure, for example, that the improvements requested by its clients 
become part of the product developed by the project. 

Analysing this from a more general point of view, this is a situation in which both parties, 
the company and the project, benefit from the collaboration. The project benefits from 
the development made by the company, or because some of its developers are paid (at 
least part-time) for their work on the project. The company benefits in knowledge about 
the product, image towards its clients, and a degree of influence over the project. 

The range of services provided by this type of company can be very broad, but 
normally consists of customised developments, adaptations or integrations 
of the products that they are experts in, or consulting services where they 
advise their clients how best to use the product in question (especially if it is 
a complex product or its correct functioning is critical for the client). 


Examples of companies that up to a point have used this business model include the 

• LinuxCare ( [45] . Established in 1996, it originally provid- 
ed consulting services and support for GNU/Linux and free software in the US, and its 
staff consisted essentially of experts in GNU/Linux. However, in 2002 its objectives 
changed and since then it has specialised in providing services almost exclusively to 
GNU/Linux running on virtual machines in large IBM computers. Its business model 
has also changed to "better knowledge with limitations", since as a fundamental part 
of its services it offers a non-free application, Levanta. 

• Alcove ( [3]. Established in 1997 in France, it mainly offers 
free software consulting services, strategic consulting, support and development. 
Since its foundation, Alcove has kept the developers of various free projects on staff, 
trying to make a return on this in terms of image. It has also tried to offer the image, 
in general, of a company linked to the free software community, by collaborating, for 
example, with user associations and giving publicity to its collaborations with free 
projects (for example, through Alcove-Labs [4]). 

5.2.2. Better knowledge with limitations 

These models are similar to those described in the previous section, but try to 
limit the competition that they may have to face. Whereas in the pure models 
based on better knowledge, anyone can, in principle, join the competition, 
since the software used is the same (and free), in this case the attempt is to 
avoid that situation by placing barriers to competition. These barriers tend to 
consist of patents or proprietary licences, which normally affect a small (but 

GNUFDL" PID_001 48386 76 Free Software 

fundamental) part of the developed product. This is why these models may be 
considered as mixed, in the sense that they are halfway between free software 
and proprietary software. 

In many cases, the free software community develops its own version, mean- 
ing that the competitive advantage can disappear, or even turn against the 
company in question if the free competitor becomes the market standard and 
is demanded by the company's own clients. 


There are many cases that use this business model, since it is frequently considered 
less risky than the pure knowledge one. However, the companies that have used it have 
evolved in different ways. Some of them include the following: 

• Caldera ( [16]. Caldera's history is complicated. In the begin- 
ning, it created its own distribution of GNU/Linux, aimed at businesses: Caldera 
OpenLinux. In 2001 it bought the Unix division from SCO, and in 2002 it changed 
its name to SCO Group. Its business strategy has changed as frequently as its name, 
from its total support for GNU/Linux, to its legal suits against IBM and Red Hat in 
2003 and abandoning its own distribution. But in relation to this heading, Caldera's 
business, at least until 2002, is a clear model of better knowledge with limitations. 
Caldera tried to exploit its knowledge of the GNU/Linux platform, but limiting the 
competition it could have faced by including proprietary software in its distribution. 
This made it difficult for its clients to change distribution once they had adopted it, 
because even though the other distributions of GNU/Linux included the free part of 
Caldera OpenLinux, they did not include the proprietary part. 

• Ximian ( [74]. Founded in 1999 under the name Helix Code by 
developers closely connected to the GNOME project, it was acquired in August 2003 
by Novell. Most of the software that it has developed has been free (in general, part 
of GNOME). However, in a very specific sphere Ximian decided to licence a compo- 
nent as proprietary software: the Connector for Exchange. This module allows one 
of its star products, Evolution (a personal information manager that includes e-mail, 
agenda, calendar, etc.,), to interact with Microsoft Exchange servers, which are com- 
monly used by large organisations. This is how it tried to compete with an advan- 
tage over the other companies that offered services based on GNOME, perhaps with 
the products developed by Ximian itself but that could not interact as easily with 
Exchange. With the exception of this product, the Ximian model has been the one of 
"better knowledge", and has also been based on being the source of a program (as we 
will see later on). In any case, this component was released as free software in 2005. 

5.2.3. Source of a free software product 

This model is similar to the one based on better knowledge but with a special- 
isation, meaning that the company using it is the producer, almost integrally, 
of a free product. Naturally, the competitive advantage increases through be- 
ing the developers of the product in question, controlling its evolution and 
having it before the competition. All of this positions the development com- 
pany very strongly towards clients who are seeking services for that program. 
Also, it is a very interesting model in terms of image, since the company has 
proven its development potential by creating and maintaining the application 
in question, which can be very useful when it comes to convincing clients of 
the company's capabilities. Likewise, it creates a good image towards the free 
software community in general, since it receives a new free product from the 
company that becomes part of the common domain. 

GNUFDL" PID_001 48386 77 Free Software 


Many free products started to be developed in a company, and very often that company 
has continued to guide its subsequent development. Some examples: 

• Ximian. We have already mentioned how it has partly used the model of better 
knowledge with limitations. But in general, Ximian has followed the clear model 
based on being the source of free programs. Its main products, like Evolution or Red 
Carpet, have been distributed under GPL licences. However, other also important 
ones, such as Mono, are distributed mostly under the MIT XI 1 or LGPL licences. In 
any case, Ximian has developed the products almost exclusively from the start. The 
company has tried to make a return on these developments by obtaining contracts 
to make them evolve in certain ways, adapting them to clients' needs, and offering 
customisation and maintenance. 

• Zope Corporation ( [75]. In 1995 Digital Creations was estab- 
lished, developing a proprietary product for the management of classified ads on the 
web. In 1997 it received a capital injection from, among others, a venture capital 
company called Opticality Ventures. What was strange about this investment (at that 
time) was the condition that was imposed of distributing the evolved product as free 
software, which later became Zope, one of the most popular content managers on 
the Internet. Since then, the company's business model has been to produce Zope 
and related products, and to offer adaptation and maintenance services for all of 
them. Zope Corporation has also known how to create a dynamic community of free 
software developers around its products and to collaborate actively with them. 

5.2.4. Product source with limitations 

This model is similar to the previous one, but takes measures to limit the com- 
petition or to maximise revenue. Among the most common limitations, we 
can find the following: 

• Proprietary distribution for a time, then released as free software. With 
or without a promise of a later free distribution, each new version of the 
product is sold as proprietary software. After a certain amount of time 
(normally, when a new version is released, also as proprietary software), 
the old version is distributed with a free licence. This way, the production 
company obtains revenue from clients interested in having the new ver- 
sions, and at the same time limits the competition, since any company 
wanting to compete using that product can only do so with the free ver- 
sion (only available when the new proprietary version is released, which 
is supposedly improved and more complete). 

• Limited distribution for a period. In this case, the software is free as of 
the moment it is first distributed. But because there is nothing in the free 
licence forcing to distribute the program to anyone who wants it (this 
is something that the person in possession of the software may or may 
not do), the producer distributes for a time to its clients only, who pay 
for it (normally in the form of a maintenance contract). After a while, it 
distributes it to anyone, for example by placing it in a public access file. 
This way, the producer obtains income from its clients, who perceive this 
preferential availability of the software as an added value. Naturally, the 
model only works if the clients do not in turn make the program public 
when they receive it. For certain types of clients, this may not be common. 

GNUFDL» PID_001 48386 78 Free Software 

In general, in these cases the development companies obtain the mentioned 
benefits, but not at zero cost. Because of the delay with which the product 
is available for the free software community, it is practically impossible for it 
to be able to contribute to its development, meaning that the producer will 
benefit very little from external contributions. 


Some companies that use this business model are as follows: 

• artofcode LLC ( [9]. Since the year 2000, artofcode sells 
Ghostscript in three versions (previously Alladin Enterprises had done this with a 
similar model). The latest version is distributed as AFPL Ghostscript, under a pro- 
prietary licence (which allows use and non-commercial distribution). The next one 
(with a year's delay more or less) is distributed as GNU Ghostscript, under the GNU 
GPL. For example, in summer 2003, the AFPL version is 8.11 (released on 16* Au- 
gust), while the GNU version is 7.07 (distributed as such on 17* May, but whose 
equivalent AFPL version is dated 2002). Also, artofcode offers a third version, with a 
proprietary licence that allows its integration with products not compatible with the 
GNU GPL (in this case it uses a dual model, which we will describe later on). 

• Ada Core Technologies ( [2]. It was established in 1994 by the 
authors of the first Ada 95 compiler, developed with partial funding from the US 
Government, based on GCC, the GNU compiler. Since the beginning its products 
have been free software. But most of them are first offered to their clients, as part 
of a maintenance contract. For example, its compiler, which continues to be based 
on GCC and is distributed under the GNU GPL, is offered to its clients as GNAT 
Pro. Ada Core Technologies does not offer this compiler to the general public by any 
means, and normally you cannot find versions of it on the Net. However, with a 
varying delay (of about one year), Ada Core Technologies offers the public versions of 
its compiler, very similar but without any type of support, in an anonymous FTP file. 

5.2.5. Special licences 

Under these models, the company produces a product that it distributes un- 
der two or more licences. At least one of them is free software, but the oth- 
ers are typically proprietary and allow the product to be sold in a more or 
less traditional way. Normally, these sales are complemented with the sale of 
consulting services and developments related to the product. For example, a 
company can distribute a product as free software under the GNU GPL, but 
also offer a proprietary version (simultaneously, and with no delay between 
the two versions) for those not wanting the conditions of the GPL, for exam- 
ple, because they want to integrate the product with a proprietary one (which 
the GPL does not allow). 


Sleepycat Software ( [60]. This 
company was established in 1996 and has announced that it has made a profit from 
the start (which is certainly remarkable in a software- related company). Its products, 
including Berkeley DB (a very popular data manager because it can be easily embedded 
in other applications), are distributed under a free licence that specifies that in the case 
of embedding with another product, it has to provide the source code of both. Sleepycat 
offers consulting and development services for its products, but also offers them under 
licences that allow them to be embedded without having to distribute the source code. 
Of course, it does this under a specific contract, and in general, under a proprietary soft- 
ware sales regime. In 2005, Sleepycat Software was bought by Oracle. 

GNUFDL" PID_001 48386 79 Free Software 

5.2.6. Brand sale 

Although it is possible to obtain very similar products for far less money, many 
clients are prepared to pay extra to buy a brand. This principle is adopted by 
companies that invest in establishing a brand with a good and well-recognised 
image that allows them to then sell free products with a sufficient margin. 
In many cases, they do not just sell those products, but also services that the 
clients will also accept as an added value. 

The most well-known cases of this business model are the companies that 
sell GNU/Linux distributions. These companies try to sell something that in 
general can be obtained at a far lower cost from the Net (or others sources 
with less of a brand image). Therefore, they have to make consumers recognise 
their brand and be prepared to pay the additional cost. To do so, they don't 
just invest in publicity, they also offer objective advantages (for example, a 
well-assembled distribution or a distribution channel that offers proximity to 
the client). Also, they tend to offer a large number of services around it (from 
training to third party certification programs), in order to make the most of 
the brand image. 


Red Hat ( [56]. Red Hat Linux started to be distributed in 1994 
(the company started to be known by its current name in 1995). For a long time, Red Hat 
managed to establish its name as the GNU/Linux distribution par excellence (although 
in the mid 2000 it shares that position with other companies like OpenSUSE, Ubuntu, 
and perhaps Debian). Several years down the line Red Hat sells all types of services related 
to the distribution, GNU/Linux and free software in general. 

5.3. Other business model classifications 

Free software literature provides other classifications of traditional business 
models. As an example, here are a few. 

5.3.1. Hecker classification 

The classification provided in "Setting up shop: the business of open source 
software" (Frank Hecker, 1998) [141] was most used in the publicity of the 
Open Source Initiative, and also one of the first to try and classify the business- 
es that were emerging around that time. However, it includes various models 
that have little to do with free software (where free software is little more than 
a companion to the main model). In any case, the models it describes are as 

GNUFDL' PID 00148386 


Free Software 

Support seller (sale of services related to the product). The company pro- 
motes a free software product (which it has developed or in which it par- 
ticipates actively) and sells services such as consulting or adaptation to 
specific requirements. 

Loss leader (sale of other proprietary products). In this case, the free pro- 
gram is used to somehow promote the sale of other proprietary products 
related to it. 


Readers will have observed 
that this classification is fairly 
different to the one that we 
have given, but even so some 
of the categories almost totally 
match some of ours. 

Widget frosting (Sale of hardware). The main business is the sale of hard- 
ware and the free software is considered a complement that can help the 
company obtain a competitive advantage. 

Accessorising (sale of accessories). Products related to free software are sold, 
such as books, computer devices, etc. 

Service enabler (sale of services). The free software serves to create a service 
(normally accessible online) from which the company makes a profit. 

Brand licensing (sale of a brand). A company registers trademarks that it 
manages to associate with free software programs, probably that it has 
developed itself. Then it obtains income through licensing the use of those 

Sell it, free it. This is a similar model to the loss leader, but done in a cycli- 
cal fashion. First a product is marketed as free software. If it is relative- 
ly successful, the next version is distributed as proprietary software for a 
time, after which it is freed. By then, a new proprietary version is being 
distributed, and so on successively. 

Software franchising. A company franchises the use of its brands in relation 
to a particular free program. 

5.4. Impact on monopoly situations 

The software market tends towards the domination of one product in each 
of its segments. Users want to make the most of the effort made in learning 
how a program works, companies want to recruit people who are familiar with 
the use of their software, and everyone wants the data that they handle to 
be manageable by the programs of the companies and people with whom 
they work. This is why any initiative designed to break a de facto situation in 
which one product clearly dominates the market is destined to produce more 
of the same: if it is successful, the new product will come to take its place, and 

GNUFDL" PID_001 48386 81 Free Software 

in a short period we will have a new dominant product. Only technological 
changes produce, during a short period, sufficient instability for nobody to 
dominate clearly 

But the fact that there is a dominant product does not necessarily have to 
lead to the creation of a business monopoly For example, petrol is a product 
that almost dominates the fuel market for private cars, but (in a free petrol 
market) there are many production companies and distribution companies for 
that same product. In reality, when we talk about software, what is worrying 
is what happens when a product manages to dominate the market because 
that product has a sole possible supplier. Free software offers an alternative 
to that situation: free products can be promoted by a specific company, but 
that company does not control them, or at least not to the extent that propri- 
etary software has us accustomed to. In the world of free software, a dominant 
product does not necessarily entail the monopoly of one company. On the 
contrary, irrespective of the product that dominates the market, many com- 
panies can compete in providing it, improving it, adapting it to clients' needs 
and offering services related to it. 

5.4.1. Elements that favour dominant products 

In computer software, it is common to have a clearly dominant product in 
each market segment. And this is normal for several reasons, among which 
we would highlight the following: 

• Data formats. In many cases the data format is very closely linked to an 
application. When a sufficiently high number of people uses it, the data 
format becomes the de facto standard, and the pressures to adopt it (and 
therefore, the application) are tremendous. 

• Distribution chains. Normally, one of the problems with starting to use 
a program is obtaining a copy of it. And it is normally difficult to find 
programs that are not leaders in their market. Distribution chains are ex- 
pensive to maintain, meaning that it is difficult for minority competitors 
to reach the computer shop where the end user can buy them. However, 
for the dominant product it is easy: the first to be interested in having it 
will be the computer shop itself. 

• Marketing. The "free" marketing that a product obtains once a significant 
proportion of the population uses it is enormous. "Word of mouth" also 
works very well when we ask and exchange information with the people 
we know. But above all the impact from the media is enormous: computer 
magazines will refer time and again to a product if it appears to be the one 
used the most; there will be training courses around it, books describing 
it, interviews with users, etc. 

GNUFDL» PID_001 48386 82 Free Software 

• Investment in training. Once time and money has been spent on learning 
how a tool functions, there is a high motivation not to change that tool. 
Also, that tool is usually the one that already dominates the market, be- 
cause it is easier to find people and materials to help teach how to use it. 

• Pre-installed software. Receiving a machine with pre-installed software is 
certainly a great incentive towards using it, even if it has to be paid for 
separately. And normally, the type of software that the seller of the ma- 
chine will be prepared to pre-install will only be the most used. 

5.4.2. The world of proprietary software 

In the world of proprietary software the appearance of a dominant product 
in any segment is equivalent to a monopoly on the part of the company that 
produces it. For example, we have these de facto monopoly situations (or al- 
most) of a product and a company in the market for operating systems, desk- 
top publishing, databases, graphic design, text processors, spreadsheets, etc. 

And this is so because the company in question has enormous control over 
the leading product, so much so that only they can drive its evolution, the 
fundamental lines along which it will be developed, its quality, etc. Users have 
very little control, since they have very little motivation to consider other 
products (for the reasons we have mentioned in the preceding section). In 
view of this, there is little that competition can do, except to try and defy the 
product's dominant position by improving their own products, (to try and 
counteract those very reasons), normally with limited success. 

This situation places the entire sector in the hands of the dominant company's 
strategy. All of the actors depend on it, and even the development of software 
technology in that field will be mediatised for the improvements that it makes 
to its product. In general terms, this is a situation where the worst economic 
effects of a monopoly arise, and in particular, the lack of motivation for the 
dominant company to tailor products to the (always evolving) needs of its 
clients, as they have become a captive market. 

5.4.3. The situation with free software 

However, in the case of free software a dominant product does not automati- 
cally translate into a business monopoly. If the product is free, any company 
can work with it, improve on it, adapt it to clients' needs, and in general, help 
it to evolve. Also, precisely due to its dominant position, there will be many 
companies interested in working with it. If the "original" producer (the com- 
pany that originally developed the product) wishes to remain in the business, 
it will have to compete with all of them and will therefore be highly motivated 

GNUFDL» PID_001 48386 83 Free Software 

to make its product evolve precisely along the lines that users want. Of course, 
it will have the advantage of better knowledge of the program, but that isn't 
all. They will have to compete for every client. 

Therefore, the appearance of dominant products in the world of free software, 
translates into more competition between companies. And with it users re- 
cover control: companies in competition cannot do anything but listen to 
them if they want to survive. And this is precisely what will make sure that 
the product improves. 

Free products that are dominant in their sector 

For a long time, Apache has been the leader in the market for web servers. But there are 
many companies behind Apache, from some very large ones (like IBM) to other much 
smaller ones. And all of them have no other choice but to compete by improving it and 
normally by contributing to the project with their improvements. Despite the fact that 
Apache is almost a monopoly in many fields (for example, it is almost the only web 
server considered on the GNU/Linux or *BSD platform), it does not depend on a single 
company, but rather on literally dozens of them. 

The distributions of GNU/Linux are also an interesting case. GNU/Linux is not, certainly, 
a monopoly, but is possibly the second choice in the market for operating systems. And 
this has not forced a situation whereby one company has control over it. On the contrary 
there are tens of distributions made by different companies, which freely compete in the 
market. Each one of them tries to offer improvements, which its competitors have to 
adopt at the risk of being left out. Moreover, they cannot stray too far from what is the 
"GNU/Linux standard", since this would be rejected by users as a "departure from the 
norm". The situation after several years of a growing market share for GNU/Linux shows 
us tens of companies that compete and allow the system to evolve. And once again, all 
of them pursue satisfying users' requirements. This is the only way that they can stay 
in the market. 

GCC is a dominant product in the world of C and C++ compilers for the GNU/Linux 
market. And yet, this has not led to any company monopoly situation, even though 
Cygnus (now Red Hat) was responsible for a long time for coordinating its development. 
There are many companies that make improvements to the system and all of them com- 
pete, each in their specific niche, to satisfy their users' demands. In fact, when a specif- 
ic company or organisation has failed in the task of coordinating (or some users have 
perceived this to be the case) there has been room for the project to fork, with two prod- 
ucts running in parallel for a while, until they have come back together again (as is now 
happening with GCC 3.x). 

5.4.4. Strategies for becoming a monopoly with free software 

Despite the fact that the world of free software is much more hostile to busi- 
ness monopolies than the world of proprietary software, there are strategies 
that a company can use to try to approach a situation of monopolistic domi- 
nance of a market. These practices are common in many other economic sec- 
tors and in order to prevent them we have bodies that regulate competition, 
which is why we will not go into too much detail about them. However, we 
will mention one that, up to a point, is specific to the software market, and 
which has already been experienced in certain situations: the acceptance of 
third party product certification. 

When a company wishes to distribute a software product (free or proprietary) 
that functions in combination with others, it is common to "certify" that prod- 
uct for a certain combination. The manufacturer undertakes to offer services 
(updates, support, problem-solving, etc.) only if the client guarantees that the 

GNUFDL" PID_001 48386 84 Free Software 

product is being used in a certified environment. For example, the manufac- 
turer of a database management program can certify its product for a certain 
GNU/Linux distribution, and no other. This implies that its clients will have 
to use that GNU/Linux distribution or forget having the manufacturer's sup- 
port (which, if the product is proprietary may be impossible in practice). If 
a particular manufacturer manages to achieve a clearly dominant position as 
a third-party certified product, users are not going to have any other choice 
than to use that product. If in that segment certification is important, we will 
once again be facing a business monopoly situation. 


Up to a point, in the market for GNU/Linux distributions we are starting to see a few cases 
of situations tending towards a de facto monopoly through certification. For example, 
there are many manufacturers of proprietary products that only certify those products on 
a given GNU/Linux distribution (very commonly Red Hat Linux). For the time being this 
is not resulting in a monopoly situation for any company, which may be due to the fact 
that certification is not so relevant for users in the market for GNU/Linux distributions. 
But only the future will tell if at some point this situation approaches a de facto monopoly. 

Nonetheless, it is important to bear in mind two comments in relation to the 
above. The first is that these monopoly positions will not be easy to achieve, 
and in any case will be achieved through "non-software" mechanisms in gen- 
eral (unlike the dominant product situation, which as we have seen is rela- 
tively normal, reached through mechanisms purely related to IT and its pat- 
terns of use). The second is that if all the software used is free, that strategy 
has limited chances of succeeding (if any at all). A manufacturer may manage 
to get lots of companies to certify for its products, but clients will always be 
able to look to different companies for services and support other than those 
that have certified for it, if they consider it appropriate. 

GNUFDL» PID 00148386 85 Free Software 

6. Free software and public administrations 

"[...] for software to be acceptable for the State, it does not only need to be technically 
capable of performing a task, but also its contracting conditions need to meet a series 
of requirements regarding licensing, without which the State cannot guarantee to its 
citizens that their data is being adequately processed, with due regard for confidentiality 
and accessibility over time, because these are highly critical aspects of its normal duty." 

Edgar David Villanueva Nunez (letter of reply to the general manager of Microsoft Peru, 

Public institutions, both those with the capacity to legislate and those dedi- 
cated to administrating the State (the "public administrations"), play a very 
important role where adopting and promoting technologies is concerned. Al- 
though until the year 2000 these institutions showed practically no interest in 
the free software phenomenon (with some exceptions), the situation started 
changing as of then. On the one hand, many public administrations started 
using free software as part of their IT infrastructure. On the other hand, in 
their role as promoters of the information society, some started to promote 
directly or indirectly the development and use of free software. Also, some 
legislative bodies have started paying attention (bit by bit) to free software, 
sometimes favouring its development, sometimes impeding it, and sometimes 
just taking its presence into consideration. 

Before going into detail, it is important to remember that for a long time free 
software was developed without explicit backing (or even interest) from public 
institutions. For this reason, the recent attention that it is drawing from many 
of them is not without controversy, confusion and problems. Also, in the last 
few years initiatives related to open standards are gaining momentum, often 
resulting in measures (more or less directly) associated to free software. 

In this chapter we will try to describe the current situation and the peculiarities 
of free software in relation to the "public" sphere. 

6.1. Impact on the public administrations 

Several studies have been made focusing on the use of free software in pub- 
lic administrations (for example, "Open source software for the public admin- 
istration", 2004 [159]; "Open source software in e-Government, analysis and 
recommendations drawn up by a working group under the Danish board of 
technology", 2002 [180]; "Free software / open source: information society op- 
portunities for Europe?", 1999 [132], and "The case for government promotion 
of open source software", 1999 [213]). Next, we will discuss some of the most 
notable ones (both positive and negative). 

GNUFDL» PID_001 48386 86 Free Software 

6.1.1. Advantages and positive implications 

Some of the advantages of using free software in public administrations and 
the main new prospects that it offers are as follows: 

1) Developing local industry 

One of the major advantages of free software is the possibility of developing a 
local software industry. When we use proprietary software, everything spent 
on the licences goes directly to the product's manufacturer, and the purchase 
strengthens the manufacturer's position, which is not necessarily negative, 
but is not very efficient for the region to which the public administration is 
associated when we consider the alternative of using a free program. 

In this case, local companies will be able to compete in providing services (and 
the program itself) to the public administration, under very similar conditions 
to any other company. Let's say that somehow the public administration is 
levelling the playing field and making it easier for anyone to compete on it. 
And of course, that "anyone" includes local companies, who will have the 
opportunity to exploit their competitive advantages (better knowledge of the 
client's needs, geographical proximity, etc.). 

2) Independence from a supplier and market competition 

Obviously, any organisation will prefer to depend on a competitive market 
than on a single provider capable of imposing the conditions under which 
it supplies its product. However, in the world of the public administration, 
this preference becomes a basic requirement, and even a legal obligation in 
some cases. In general, the public administration cannot choose to contract 
a given supplier, but rather must specify its requirements in such a way that 
any interested company that fulfils certain characteristics and that offers the 
required product or service, can opt for a contract. 

Once again, in the case of proprietary software, each product has just one 
supplier (even if it uses a number of intermediaries). If a particular product is 
specified, then the public administration will also be deciding what provider 
to award the contract. And in many cases it is virtually impossible to avoid 
specifying a particular product, when we are dealing with computer programs. 
Reasons of compatibility within the organisation or savings in training and 
system administration, or many more, make it common for a public body to 
decide to use a certain product. 

The only way out of this situation is by making the specified product free. 
This way, any interested company will be able to supply it and also any type 
of service related to it (subject only to the company's capabilities and knowl- 
edge of the product). Also, in the case of this type of contracting, the public 

GNUFDL» PID_001 48386 87 Free Software 

administration can change supplier in the future if it wishes, and without any 
technical problems, since even if it changes company, it will still be using the 
same product. 

3) Flexibility and adaptation to specific requirements 

Although adaptation to specific requirements is something that any organisa- 
tion using computers needs, the peculiarities of the Administration make this 
a very important factor in the successful implantation of a software system. 
In the case of free software, the adaptation is made much easier, and more 
importantly, can rely on a competitive market if contracting it is necessary. 

When the public administration buys a proprietary product, modifying it nor- 
mally involves reaching an agreement with the manufacturer, who is the only 
party that can legally (and often technically) do it. Under these circumstances, 
it is difficult to negotiate, especially if the manufacturer is not excessively in- 
terested in the market offered by that particular administration. However, by 
using a free product, the Administration can modify it as it wishes, if it em- 
ploys capable personnel, or outsource the modification. In principle, this out- 
sourcing is possible with any company that has the skills and knowledge to 
do so, meaning that several companies can be expected to compete. Naturally, 
this tends to make the cost cheaper and improve the quality. 

The case of GNU/Linux distributions 

In the last few years in Spain, it has become common for certain regional governments 
to create their own GNU/Linux distributions. This trend started with GNU/Linux, but 
nowadays there are many more. Although some experts have criticised the existence 
of these distributions, it is a clear example of the flexibility that free software allows. 
Any public administration, by spending relatively moderate resources, can contract a 
GNU/Linux adaptation adapted to its needs and preferences, without practically any 
limits. For example, it can change the desktop appearance, choose the set of default 
applications and language, improve the applications' localisation, etc. In other words: if 
wanted, the desktop (and any other software element that works on the computer) can 
be adapted to precise requirements. 

Of course, this adaptation will involve some expenditure, but experience shows that 
it can be achieved relatively cheaply, and the trend appears to indicate that it will be 
increasingly easier (and cheaper) to make customised distributions. 

4) Easier adoption of open standards 

Given their very nature, free programs commonly use open or non-proprietary 
standards. In fact, almost by definition, any aspect of a free program that we 
may care to consider can be reproduced easily and, therefore, is not propri- 
etary. For example, the protocols used by a free program in order to interact 
with other programs can be studied and reproduced, meaning that they are 
not proprietary. But also, quite commonly and in the interest of the projects 
themselves, we try to use open standards. 

GNUFDL» PID_001 48386 88 Free Software 

In any case, irrespective of the motive, it is a fact that free programs normally 
use non-proprietary standards for data exchange. The advantages of this for 
public administrations are more far-reaching than for any other organisation, 
since the promotion of proprietary standards (even indirectly, by using them) 
is much more of a concern in their case. And in at least one aspect, the use 
of non-proprietary standards is fundamental, where interaction with citizens 
is concerned, since they should not be forced to purchase any product from 
a particular company in order to be able to interact with the public adminis- 

5) Public scrutiny of security 

For a public administration, being able to guarantee that its computer systems 
only do what they have to is a fundamental obligation, and in many countries, 
a legal requirement. Often these systems handle private data, which third par- 
ties could be interested in (for example tax data, criminal records, census or 
electoral data, etc.). If a proprietary application is used, without source code 
available, it is difficult to guarantee that the application will process the data 
in the way that it should. But even if it does provide its source code, the pos- 
sibilities of a public institution ensuring that it does not contain strange code 
will be very limited. Only if the task can be habitually and routinely commis- 
sioned to third parties, and plus any interested party can scrutinise it, can the 
Administration be reasonably sure that it is complying with its fundamental 
duty, or at least taking the measures within its power to do so. 

6) Availability in the long term 

Much of the data processed by the administrations, and the programs used 
to calculate them, need to be available within decades and decades. It is very 
difficult to guarantee that any proprietary program will be available after this 
time, especially if the idea is for it to work on the usual platform at that time 
in the future. On the contrary, it is possible that the manufacturer may have 
lost interest in the product and has not ported it to new platforms, or is only 
prepared to do so for a lot of money. Once again, we need to remember that 
only the manufacturer can port the product, meaning that negotiations will 
be difficult. In the case of free software, however, the application is available, 
with certainty, so that anyone can port it and leave it functioning according 
to the needs of the Administration. If this does not happen spontaneously, 
the Administration can always look for several companies to make the best 
offer to do the job. This guarantees that the application and the data that it 
processes will be available when needed. 

7) Impact beyond use on the part of the Administration 

GNUFDL' PID 00148386 


Free Software 

Many applications used or promoted by the public administrations are also 
used by many other sectors of society. For this reason, any public investment 
in the development of a free product benefits not only the Administration 
itself, but also all its citizens, who will be able to use the product for their 
computer tasks, perhaps with the improvements made by the Administration. 


A very particular case, but one with enormous impact, which displays this better use of 
public resources is program localisation (adaptation to a community's uses and customs). 
Although the most visible aspect of localisation is the translation of the program and 
its documentation, there are others that are also affected by it (from use of the local 
currency symbol to presenting the date and time in the formats of the community in 
question, to the use of examples in the documentation and ways of expression adapted 
to local customs). 

In any case, obviously if a public administration uses funds to localise a particular appli- 
cation tailoring the application to its needs, it is more than likely that those needs coin- 
cide with those of its citizens, meaning that it will generate, not only a program that sat- 
isfies its own requirements, but also, one that can be made available to any citizen able to 
make the most of it at no additional cost. For example, when an administration finances 
the adaptation of a computer program to a language that is used within its community, 
it will not only be able to use that program within its own offices, but also offer it to cit- 
izens, with everything that this involves in terms of developing the information society. 


Readers interested in a re- 
port on the advantages of 
free software for the Admin- 
istration, written in the US 
context of 1999, can consult 
"The case for government 
promotion of open source 
software" (Mitch Stoltz, 
1999) [213]. 

6.1.2. Difficulties of adoption and other problems 

However, although there are many advantages for the administration using 
free software, there are also many difficulties that need to be faced when it 
comes to putting it into practice. Of them, we would particularly mention the 

1) Lack of knowledge and political commitment 

The first problem that free software encounters for entering the Adminis- 
tration is one that other organisations undoubtedly share: free software is 
still unknown for the people who make the decisions. 
Fortunately, this is a problem that is gradually being solved, but in many 
spheres of the Administration, free software is still perceived as something 
strange, so decisions about using it still involve certain risks. 
In addition to this, we tend to come across a problem of political deci- 
sion-making. The main advantage of free software for the Administration 
is not the cost (since the cost, in any case, is high, especially when we are 
talking about a roll-out for a large number of workstations), but as we have 
already said, benefits are above all strategic. And therefore, the decision 
falls within the political, rather than the technical sphere. Without the 
political will to change software systems and the philosophy with which 
they are contracted, it is difficult to progress with the deployment of free 
software in the Administration. 

2) Poor adaptation of contracting mechanisms 

The contracting mechanisms that the Administration uses nowadays, 
ranging from the usual public tender models to cost itemising, are funda- 
mentally designed for the purchase of IT products and not so much for 

GNUFDL" PID_001 48386 90 Free Software 

the purchase of services related to programs. However, when we use free 
software, normally there is no product to be bought, or its price is negli- 
gible. In contrast, to take advantage of the opportunities provided by free 
software, it is convenient to be able to contract services around it. This 
makes it necessary, before free software can be seriously used, to design 
bureaucratic mechanisms that facilitate contracting in these cases. 

3) Lack of deployment strategy 

Often an administration may start to use free software simply because the 
purchase cost is lower. It is common in these cases for the product in ques- 
tion to be incorporated into the computer system with no further plan- 
ning, and in general, without a global strategy for using and making the 
most of the free software. This causes most of its benefits to be lost along 
the way, since everything boils down to the use of a cheaper product, where- 
as we have already seen that, in general, the major benefits are of a differ- 
ent type. 

If added to this, the transition is not properly designed, the use of free 
software can incur considerable costs, and we will see that in certain iso- 
lated cases, outside of a clear framework, the use of free software in the 
Administration can be unsuccessful and frustrating. 

4) Scarcity or lack of free software products in certain segments 

The deployment of free software in any organisation can encounter the 
lack of free quality alternatives for certain types of applications. For these 
cases, the solution is complicated: all that we can do is try to promote the 
appearance of the free product that we need. 

Fortunately, public administrations are in a good position to study seri- 
ously whether they may be interested in promoting or even financing or 
co-financing, the development of that product. We should remember that 
its objectives normally include providing its citizens with better access to 
the information society, for example, or promoting the local industrial 
fabric. Certainly, the creation of many free programs will have a positive 
influence on both objectives, meaning that we should add to the mere 
direct cost/benefit calculation, the indirect benefits that such a decision 
will have. 

5) Interoperability with existing systems 

It is not common for a full migration to free software to be made with 
the entire system at the same time. Therefore, it is important for the part 
that we want to migrate to continue functioning correctly in the context 
of the rest of the software with which it will have to interoperate. This 
is a well-known problem with any migration (even if it is a proprietary 
product), but it can have a particular impact in the case of free software. 
In any case, it will be something to be taken into account when study- 
ing the transition. Fortunately, we can often adapt the free software that 
needs to be installed to interoperate adequately with other systems, but if 

GNUFDL" PID_001 48386 91 Free Software 

this is needed, this point will have to be considered when budgeting the 
migration costs. 

6) Data migration 

This is a generic problem of any migration to new applications that use 
different data formats, even if they are proprietary In fact, in the case of 
free software this problem is often mitigated, since it is usual to make a 
special effort to foresee as many formats and data exchange standards as 
possible. But normally the data has to be migrated. And the cost of doing 
this is high. Therefore, in calculating the cost of a potential migration to 
free software, this factor needs to be carefully considered. 

6.2. Actions of the public administrations in the world of free 

Public administrations influence the world of software in at least three ways: 

• By buying programs and services related to them. Administrations, as large 
users of software, are fundamental players in the software market. 

• By promoting different ways of using (and purchasing) certain programs 
by individuals or companies. This promotion is sometimes achieved by of- 
fering financial incentives (tax deductions, direct incentives, etc.), some- 
times through information and advice, sometimes by "follow my exam- 

• By financing (directly or indirectly) research and development projects 
that design the future of software. 

In each of these spheres free software can offer specific advantages (in addition 
to those already described in the preceding section) of interest to both the 
Administration and to society in general. 

6.2.1. How to satisfy the needs of the public administrations? 

Public administrations are large consumers of IT. Where software is concerned, 
they normally buy off-the-shelf products as well as customised systems. From 
this point of view, they are fundamentally large purchasing centres, similar to 
those of big companies, but with their own peculiar features. For example, in 
many spheres, the purchasing decisions of the public administrations are sup- 
posed to take into consideration not only cost versus functionality parameters, 
but also others, such as the impact of the purchase on the industrial or social 
welfare or long term strategic considerations, which can also be important. 

GNUFDL» PID_001 48386 92 Free Software 

In any case, the usual nowadays with off-the-shelf software is to use market 
leader proprietary products. The amount of public money spent by municipal- 
ities, regional and national governments, and international (such as European 
Union) public administrations on purchasing Windows, Office or other simi- 
lar product licences is certainly considerable. But gradually free solutions are 
starting to enter the market. Increasingly, solutions based on free software are 
being considered for servers, and products such as OpenOffice.og, and GNU/ 
Linux with GNOME or KDE are increasingly used for the desktop. 

What is there to be gained from this migration to free software? To illustrate 
just what, let's consider the following scenario. Let's suppose that with a frac- 
tion of what is spent on two or three "star" proprietary products by all the 
European administrations (or probably those of any medium-sized developed 
country), we could convene a public tender for one company (or two, or three, 
or four) to improve and adapt the currently available free programs so that 
within one or two years they would be ready for massive use, at least for cer- 
tain standard tasks (if they are not already). Let's imagine for example, a co- 
ordinated effort, on a national or European scale, whereby all the administra- 
tions participated in a consortium responsible for managing these tenders. In 
a short period of time there would be a "local" industry specialised in making 
these improvements and adaptations. And the administrations could choose 
between the three or four free distributions produced by that industry. In or- 
der to promote competition, each company could be compensated according 
to the number of administrations that chose to use their distribution. And the 
entire result of this operation, because it would be free software, would also 
be available for companies and individual users, which in many cases would 
have similar needs to the administrations'. 

In the case of customised software, the normal process currently involves con- 
tracting the necessary programs from a company under a proprietary model. 
Any development made at the Administration's request is the property of the 
company that develops it. And usually, the contracting administration is tied 
to the supplier in everything related to improvements, updates, and support, 
in a vicious circle that makes competition difficult and slows down the pro- 
cess of modernising public administrations. Even worse is that often the same 
program is sold time and again to similar administrations, applying in each 
case the costs incurred for making the development from scratch. 

Let's consider again how things could be different. A consortium of public ad- 
ministrations needing a particular type of customised software could demand 
that the obtained result be free software. This would allow other administra- 
tions to benefit from the work and in the medium term may interest them in 
collaborating in the consortium so that their particular requirements could be 
taken into consideration. Because the resulting software would be free, there 
would be no obligation to contract the improvements and adaptations to the 
same supplier, meaning that competition would enter the market (which at 

GNUFDL» PID_001 48386 93 Free Software 

present is almost captive). In all the aforementioned situations, the final cost 
for any of the administrations involved would never be more than if a propri- 
etary model had been adopted. 

Are these scenarios science fiction? As we will see later, there are timid ini- 
tiatives in similar directions to the ones described. In addition to helping to 
create and maintain an industry within the sphere of the purchasing public 
administration, free software offers more specific advantages in the public do- 
main. For example, it is the most efficient way of having software developed 
in minority languages (a basic concern of many public administrations). It can 
also help a lot towards maintaining strategic independence in the long term 
and ensuring the accessibility of the data in public administrations' custody 
for a long time. For all of these reasons, public bodies are increasingly inter- 
ested in free software as users. 

Some cases related to German administrations 

In July 2003 the first stable version of Kolab was released, a product of the Kroupware 
project. Kolab is a free IT help system for group work (groupware) based on KDE. The 
reason for mentioning this project is that originally it was a tender by the German 
government's Bundesamt fur Sicherheit in der Informationstechnik (BSI - translated as 
the Federal Office for Information Security). This tender sought a solution that would 
interoperate with Windows and Outlook on the one hand, and GNU/Linux and KDE on 
the other. Of the submitted bids, the joint proposal of three companies, Erfrakon, Inte- 
vation and Klaralvdalens Datakonsult, was awarded the contract, with their proposal to 
provide a free solution partly based on software already developed by the KDE project, 
completed with its own free developments, resulting in Kolab. 

In May 2003, the Town Hall of Munich (Germany) approved the migration to GNU/Lin- 
ux and free office suite applications for all desktop computers, about fourteen thousand 
in total. The decision to do this was not purely financial: strategic and qualitative aspects 
were also taken into consideration, according to the authorities. In the comprehensive 
analysis that was carried out prior to making the decision, the solution that was finally 
chosen (GNU/Linux plus, fundamentally) obtained 6,218 points (from 
a maximum of ten thousand) as opposed to the little more than five thousand points 
obtained by the "traditional" solution based on Microsoft software. 

In July 2003, the Koordinierungs-und Beratungsstelle der Bundesregierung fur Informa- 
tionstechnik in der Bundesverwaltung (KBSt), under the German Ministry of the Interior, 
made public the document Leitfaden fur die Migration von Basissoftwarekomponenten auf 
Server- und Arbeitsplatzsystemen [107] ('Migration guide for the basic software components 
of servers and workstations'), which offers German public bodies a set of guidelines on 
how to migrate to solutions based on free software. These guidelines are designed for the 
decision-making party to evaluate whether a migration to free software is appropriate 
and how to carry out the migration if that decision is made. 

6.2.2. Promotion of the information society 

Public bodies spend a lot of resources on incentives to encourage IT spending. 
This is a formidable tool, which can help new technologies to expand in soci- 
ety. But it is also a dangerous tool. For example, it may not be a very good idea 
to promote society's use of the Internet by recommending a particular navi- 
gator encouraging one company's de facto monopoly position, because in the 
long term this could be negative for the society that we are trying to benefit. 

GNUFDL» PID_001 48386 94 Free Software 

Once again, free software can help in these situations. In the first place, it 
is neutral towards manufacturers, since nobody has the exclusivity over any 
free program. If an administration wishes to promote the use of a family of 
free programs, it can convene a tender, which any company in the sector can 
bid for, to manage its delivery to citizens, its improvement or extended func- 
tionality, etc. Secondly, it can help a lot in economic aspects. For example, in 
many cases the same amount of funds can be spent on purchasing a certain 
number of licences for proprietary programs as for purchasing one free copy 
and contracting support or adaptations for it; or even on negotiating with a 
proprietary software manufacturer for the rights to convert its product into 
free software. 

In a separate field, we could imagine dedicating part of the amount allocat- 
ed for the computerisation of schools to creating a GNU/Linux distribution 
adapted to primary schools' teaching requirements. And with the rest of the 
funds contracting support for maintaining the software in those schools, so 
that the software is not merely "for show" but rather people are genuinely 
responsible for ensuring that it works correctly. This not only covers educa- 
tional requirements but also generates a market for companies, usually local 
ones, capable of providing maintenance services. And of course, it leaves the 
path to the future completely open: the software will not become obsolete in 
just a few years meaning that we need to start over from scratch, rather it can 
be updated incrementally, year after year, maintaining the program's benefits 
with a similar investment. 


Readers who are familiar with public initiatives in respect of free software will recognise 
the case of gnuLinEx in this example. Towards the end of 2001, the Regional Govern- 
ment of Extremadura (Spain) decided to use a GNU/Linux distribution in order to com- 
puterise all of the public schools in the region. To do so, it financed the construction 
of gnuLinEx, a GNU/Linux distribution based on Debian that was announced in spring 
2002, and made sure that it was a requirement in all tenders for purchasing schools' 
computer equipment. Also, it started training programs for teachers, creating teaching 
materials and expanding the experience into other fields. In mid- 2003, it seemed that 
the experience was a success, as it expanded institutionally to other regions (for example, 
to Andalucia, also in Spain, through the Guadalinex project). 

6.2.3. Research promotion 

Free software also provides noteworthy benefits where R+D policies are con- 
cerned. Public money is being used to finance a large amount of software de- 
velopment that society does not end up benefiting from, even indirectly. Usu- 
ally, public research and development programs finance, wholly or in part, 
projects to create software without really worrying about the rights that the 
public will have over them. In many cases the results, without an adequate 
commercialisation plan, are simply filed and left to gather dust. In others, the 
same people who financed a program through taxes end up paying for it again 
if they wish to use it (since they need to buy licences for use). 

GNUFDL» PID_001 48386 95 Free Software 

Free software offers an interesting choice, which the authorities responsible for 
innovation policies in many administrations are gradually starting to consider 
with care. Especially when the research is pre-competitive (most common in 
the case of public funding), the fact that resulting programs are free allows 
industry as a whole (and consequently society) to benefit enormously from the 
public money spent on R+D in the software field. Where one company may 
see a result that is impossible to sell, another may see a business opportunity. 
This way, on the one hand, the results of research programs are maximised, 
and on the other, competition between companies wishing to use the results 
of a project increases, since all of them will compete on the basis of the same 
programs resulting from the project. 

This model is not new. To a great extent it is the one that has allowed the 
Internet to develop. If public administrations demand that the results of re- 
search carried out with its funds is distributed in the form of free software, it 
would not be surprising for similar cases to appear, on different levels. Either 
the outcome of that research will be poor or useless (in which case the way 
of selecting funding projects needs to be reviewed), or the dynamic generated 
by leaving them ready for any company to be able to convert them into a 
product would allow simply unforeseeable developments. 

6.3. Examples of legislative initiatives 

In the following sections we look at some specific legislative initiatives relating 
to the use and promotion of free software by public administrations. Of course, 
the list we provide does not intend to be exhaustive, and has focused on the 
initiatives that have been pioneering in some way (even if they were not fi- 
nally approved). Interested readers can complete it by consulting "GrULIC. 
Legislation regarding the use of free software by the State" [133], which cites 
many more similar cases. Also, in one appendix (appendix D) we include for 
illustrative purposes the full text or the most relevant parts of several of these 

6.3.1. Draft laws in France 

In 1999 and 2000 in France two draft laws related to free software were pre- 
sented, which were pioneers in a long series of legislative debates over the 

• Draft law of 1999-495, proposed by Laffitte, Tregouet and Cabanel, 
was made available on Senate of the French Republic's web server in 
October 1999. Following a process of public debate over the Internet 
( [102] which lasted two 
months, the draft was modified. The result was Draft Law number 2000- 
117 (Laffitte, Tregouet and Cabanel, Proposition de Loi numero 117, Sen- 
ate of the French Republic, 2000) [162], which advocated the obligatory 
use of free software by the Administration, contemplating exceptions and 

GNUFDL» PID_001 48386 96 Free Software 

transition measures for cases where it was not yet technically possible, in 
the more general context of expanding the use of the Internet and free 
software across the French administration. 

• In April 2000, members of parliament Jean-Yves Le Deaut, Christian Paul 
and Pierre Cohen proposed a new law whose objective was similar to that 
of Laffitte, Tregouet and Cabanel's draft: to reinforce the freedoms and 
security of consumers, in addition to improving the equality of rights in 
the information society 

However, unlike the draft law of Laffitte, Tregouet and Cabanel, this sec- 
ond one did not make it compulsory for the Administration to use free 
software. This draft law centred on the fact that the software used by the 
Administration should have the source code available, but without forcing 
it to be distributed with free software licences. 

In order to achieve their objectives, the legislators aimed to guarantee the 
software's "right to compatibility", by providing mechanisms that put into 
practice the principle of interoperability reflected in EC Directive related 
to the legal protection of computer programs (Council Directive 91/250/ 

EEC, of 14 l May 1991, regarding the legal protection of computer pro- 
grams, 1991) [111]. 

Neither of the two French drafts was passed into law, but both have served to 
inspire most subsequent initiatives worldwide, which is why they are particu- 
larly interesting to study. The second one (proposed by Le Deaut, Paul and Co- 
hen) pursued the compatibility and interoperability of the software, empha- 
sising the availability of the source code for the software used by the Adminis- 
tration. However, it did not require developed applications to be free software, 
understood as meaning software distributed under licences that guarantee the 
freedom to modify, use and redistribute the program. 

Later on (section D.l and section D.2 of appendix D) we reproduce almost 
in full the articles and explanatory memorandums of both draft laws. The 
explanatory memorandums are particularly interesting, as they highlight the 
problems currently threatening the public administrations regarding the use 
of software in general. 

6.3.2. Draft law of Brazil 

In 1999, parliament member Walter Pinheiro presented a draft law on free 
software to the Federal Chamber of Brazil (Proposicao pl-2269/1999. Dispoe 
sobre a utilizacao de programas abertos pelos entes de direito publico e de 
direito privado sob controle acionario da administracao publica, Chamber of 
Deputies of Brazil, December 1999) [185]. This project concerned the use of 
free software in the public administration and in private companies with the 
State as majority shareholder. 

GNUFDL» PID_001 48386 97 Free Software 

It recommends the use of free software by these bodies with no restrictions in 
terms of lending, modification or distribution. The articles of the law describe 
in detail how free software is defined and how the licences that come with it 
should be. The definitions coincide with the classical definition of free soft- 
ware by the GNU project. The explanatory memorandum reviews the history 
of the GNU project, analysing its advantages and achievements. It also refers 
to the current situation of free software, using the GNU/Linux operating sys- 
tem as an example. 

One of the most interesting parts of article one, defines very clearly the sphere 
in which the use of free software is proposed (bearing in mind that the defi- 
nition provided in later articles for "open program" is, as already mentioned, 
the same as free software): 

"The Public Administration at all levels, the powers of the Republic, State and mixed 
public-private enterprises, public companies and all other public or private bodies sub- 
ject to the control of the Brazilian State are obliged to use preferably, in their computer 
systems and equipment, open programs, free of proprietary restrictions with regards to 
their cession, modification and distribution." 

6.3.3. Draft laws in Peru 

In Peru, several draft projects have been proposed on the use of free software by 
the public administration ("GNU Peru. Draft laws on free software in the public 
administration of the Peruvian government", Congress of the Republic) [184]. 
The first and most renowned was proposed by congressman Edgar Villanueva 
Nunez in December 2001 (Draft law on free software number 1609, December 
2001) [222]. It defines free software according to the classical definition of the 
four freedoms (adding perhaps more legal precision, with a definition that 
specifies six characteristics to be a free program) and proposes its exclusive use 
in the Peruvian administration: 

"Article 2. The executive, legislative and judicial authorities, decentralised bodies and 
companies where the State is the majority shareholder, will use exclusively free programs 
or software in their computer systems and equipment." 

Nevertheless, later on, articles 4 and 5 include certain exceptions to this rule. 

In its day this draft law had a global repercussion. On the one hand, it was 
the first time that an administration's exclusive use of free software had been 
proposed. But even more importantly for the repercussion of this novelty, was 
the epistolary exchange between congressman Villanueva and Microsoft's rep- 
resentation in Peru, which made allegations against the proposal. This draft 
law is also interesting in relation to the position adopted by the US embassy, 
which even sent through official channels a notification (attaching a report 
prepared by Microsoft) to the Peruvian Congress expressing its "concern over 
recent proposals by the Congress of the Republic to restrict purchases of the 
Peruvian Government to open source software or free software " ("Letter to the 
president of the Congress of the Republic", 2002) [147]. Among other motives, 
the allegations of both Microsoft and the US Embassy tried to prove that the 

GNUFDL» PID_001 48386 98 Free Software 

draft law would discriminate between different companies making impossible 
the investments required in order to generate a national industry of software 
creation. Villanueva argued back that the draft law did not discriminate or 
favour any particular company in any way, since it did not specify who the 
Administrator's supplier could be, but rather how (in what conditions) the 
software would have to be provided. This reasoning is very clear for under- 
standing how the Administration's promotion of free software does not in any 
way prejudice free competition between providers. 

Later on, Peruvian congressmen Edgar Villanueva Nunez and Jacques Rodrich 

Ackerman presented a new draft law, number 2485, of 8 th April 2002 (Draft 
Law on the Use of Free Software in the Public Administration number 2485, 
2002) [223], which in August 2003 was still in parliamentary proceedings. This 
draft law was an evolution of Draft Law 1609 [222], from which it draws sev- 
eral comments making several improvements, and may be considered a good 
example of a draft law that proposes the exclusive use of free software in the 
public administrations, save for certain exceptional cases. Given its relevance, 
we include its full text (section D.3 of appendix D). In particular, its explana- 
tory memorandum is a good summary of the characteristics that the software 
used by the public administrations should have and how free software com- 
plies with these characteristics better than proprietary software. 

6.3.4. Draft laws in Spain 

In Spain there have been several legislative initiatives related to free software. 
Below, we cite a few of them: 

• Decree of Measures to Promote the Knowledge Society in Andalucia 
One of the most important legislative initiatives in Spain (because it has 
come into force) has been unquestionably the one adopted by Andalucia. 
The Decree of Measures to Promote the Knowledge Society in Andalucia 

(Decree 72/2003, of 18 l March of Measures to Promote the Knowledge 
Society in Andalucia, 2003) [99] deals with the use of free software, fun- 
damentally (but not only) in the educational context. 
Among others, it promotes the preferable use of free software in public ed- 
ucational centres, obliging all of the equipment purchased by these cen- 
tres to be compatible with free operating systems, and likewise for the Re- 
gional Government centres that provide public Internet access. 

• Draft law on Free Software in the context of the Public Administration of 

Other communities have debated more ambitious proposals, but without 
obtaining the majority vote that they required. The most renowned of 
them is probably the one debated in the Parliament of Catalonia (Proposi- 
cio de llei de programari lliure en el marc de l'Administracio publica de 
Catalunya, 2002) [221], very similar to the one that the same party (Es- 

GNUFDL» PID_001 48386 99 Free Software 

querra Republicana de Catalunya) presented to the Congress of Deputies, 
which we will talk about next. This proposal was unsuccessful when sub- 
mitted for voting. 

• Draft Law of Puigcercos Boixassa in the Congress of Deputies 

There was also an initiative in the Congress of Deputies proposed by Joan 
Puigcercos Boixassa (Esquerra Republicana de Catalunya) (Draft Law of 
Measures for the Implantation of Free Software in the State Administra- 
tion, 2002) [188]. This initiative proposed the preferable use of free soft- 
ware by the State Administration, and in this sense is similar to other ini- 
tiatives that share this objective. However, it has the interesting peculiari- 
ty of emphasising the availability of localised free programs for the co-of- 
ficial languages (in the autonomous communities that have them). The 
initiative was not approved in parliamentary proceedings. 

GNUFDL' PID 00148386 


Free Software 

7. Free software engineering 

"The best way to have a good idea is to have many of them." 
Linus Pauling 

In previous chapters we have shown why free software's challenge is not the 
one of a competitor that generates software more quickly, more cheaply and 
of better quality: free software is different from "traditional" software in more 
fundamental aspects, starting with philosophical reasons and motivations, 
continuing with new market and economic models, and ending with a dif- 
ferent way of generating software. Software engineering cannot be unaffected 
by all of the aforementioned factors; so, for a little more than over ten years 
research on how free software is developed has been targeted in greater depth. 
This chapter aims to discuss the most significant studies and the evidence that 
they provide, with a view to offering the reader a vision of the state of the 
art and the future prospects of what we have decided to call free software en- 

7.1. Introduction 

Although free software has been developed for several decades now, it is on- 
ly in recent years that we have started to pay attention to its development 
models and processes from a software engineering perspective. In the same 
way as there is no single model for proprietary software development, there 
is no single model for free software 5 development, but even so we can find 
interesting characteristics that most of the projects under study share and that 
could stem from the properties of free programs. 

( T"he article "The ecology of open 
source software development" 
(Kieran Healy and Alan Schussman, 
2003) [140] shows the large vari- 
ety of projects and their diversity 
in numbers of developers, use of 
tools and downloads. 

In 1997, Eric S. Raymond published the first broadly disseminated article The 
cathedral and the bazaar. Musings on Linux and open source by an accidental rev- 
olutionary, O'Reilly & Associates, 2001) [192], describing 
some of the characteristics of free software development models, making spe- 
cial emphasis on what distinguished these models from those of proprietary 
development. Since then, this article has become one of the most renowned 
(and criticised) in the world of free software, and up to a point, the sign of the 
starting development of free software engineering. 

7.2. The cathedral and the bazaar 

Raymond makes an analogy between the way of building mediaeval cathedrals 
and the classical way of producing software. Arguing that in both cases there 
is a clear distribution of tasks and functions, emphasising the existence of a 
designer who oversees everything and has to control the development of the 

GNUFDL» PID 00148386 101 Free Software 

activity at all times. At the same time, planning is strictly controlled, giving 
rise to detailed processes where ideally each participant in the activity has a 
clearlv defined role. 

clearly defined role. 

What Raymond takes as the model for building cathedrals not only has room 
for the heavy processes that we can find in the software industry (the classical 
waterfall model, the different aspects of the Rational Unified Process, etc.), 
but also for free software projects such as GNU and NetBSD. For Raymond, 
these projects are highly centralised, since just a few people are responsible 
for the software's design and implementation. The tasks carried out by these 
people, in addition to their functions, are well defined, and anyone wishing 
to form part of the development team needs to be assigned a task and a func- 
tion according to the project's requirements. On the other hand, releases of 
this type of programs are spaced in time according to a fairly strict schedule. 
This means having few software releases and long cycles, consisting of several 
stages between releases. 

The opposite model to the cathedral is that of the bazaar. According to Ray- 
mond, some of the free software programs, particularly the Linux kernel, have 
been developed following a similar scheme to that of an oriental bazaar. In 
a bazaar there is no maximum authority to control the processes that are de- 
veloped or to strictly plan what has to happen. At the same time, participants' 
roles can change continuously (sellers can become clients) and with no out- 
ward indication. 

But what is most novel about "The cathedral and the bazaar" is how it describes 
the process by which Linux has become a success; it is a list of "good practices" 
to make the most of the opportunities offered by the source code being avail- 
able, and of interactivity through the use of telematic systems and tools. 

A free software project tends to appear as a result of a purely personal action; 
the cause can be found in a developer who finds his ability to resolve a problem 
limited. The developer needs to have enough knowledge to start solving it, 
at least. Once he has obtained something usable, with some functionality, 
simple, and if possible, well designed or written, the best he can do is to share 
that solution with the world of free software. This is what is known as release 
early, which helps to draw the attention of other people (usually developers) 
who have the same problem and who may be interested in the solution. 

One of the basic principles of this development model is to think of users as 
co-developers. They need to be treated with care, not only because they can 
provide "word of mouth" publicity but also because they will carry out one of 
the most costly tasks that there is in software generation: testing. Unlike co- 
development, which is difficult to scale, debugging and tests have the proper- 
ty of being highly parallelisable. The user will be the one to take the software 

GNUFDL» PID_001 48386 102 Free Software 

and to test it on his machine under specific conditions (an architecture, cer- 
tain tools, etc.), a task that multiplied by a large number of architectures and 
environments would entail an enormous effort for the development team. 

If we treat users as co-developers it could happen that one of them finds a bug 
and resolves it by sending a patch to the project developers so that the problem 
can be solved in the following version. It can also happen, for example, that 
someone other than the person who discovers the bug eventually understands 
it and corrects it. In any case, all of these circumstances are beneficial for the 
development of free software, i.e. it is beneficial to enter a dynamic of this 

This situation becomes more effective with frequent releases, since the moti- 
vation to find, notify and correct bugs is high because it is assumed that they 
will be attended immediately. Also, secondary benefits are achieved such as 
the fact that frequent integration - ideally once or more times a day - does 
not require a final phase of integrating the modules comprising the program. 
This has been called release often and allows a great modularity (Alessandro 
Narduzzo and Alessandro Rossi, "Modularity in action: GNU/Linux and free/ 
open source software development model unleashed", May 2003) [176], at the 
same time as it maximises the propaganda effect provided by the publication 
of the software's latest version. 


New version management depends, logically, on the size of the project, since the prob- 
lems that need to be dealt with are not the same when the development team has two 
members as when it has hundreds. Whereas, in general, for small projects this process 
is more or less informal, the management of releases for large projects tends to follow a 
defined process, which is not exempt from a certain degree of complexity. There is an 
article called "Release management within open source projects" (Justin R. Ehrenkrantz, 
2003) [110] which describes in detail the sequence followed with the Apache web server, 
the Linux kernel and the Subversion versioning system. 

In order to prevent "release often" from frightening users with a priority for the 
stability of the software over the speed with which the software evolves, some 
free software projects have several development branches running in parallel. 
The most renowned case of this is the Linux kernel, which historically has had 
directed at those who value its reliability and another unstable one designed 
for developers with the latest innovations and novelties. 

7.3. Leadership and decision-making in the bazaar 

Raymond suggests that all free software projects should have a benevolent dic- 
tator, a sort of leader who is normally the founder of the project to guide the 
project and always have the last word when it comes to decision-making. The 
skills that this person must have involve mainly knowing how to motivate 
and coordinate a project, understanding users and co-developers, seeking con- 

GNUFDL» PID_001 48386 103 Free Software 

sensus and integrating everyone who has something to contribute. As you can 
see, we have not mentioned technical competence among the most important 
requirements, although it is never superfluous. 

As the size of projects and the number of developers involved with them have 
grown, new ways of organising decision-making have emerged. Linux, for ex- 
ample, has a hierarchical structure based on Linus Torvalds delegating respon- 
sibilities, the "benevolent dictator". And, we will see that there are parts of 
Linux that have their own "benevolent dictators", although their power will 
be limited by the fact that Linus Torvalds has the last word. This case is a clear 
example of how a high level of modularity in a free software project has given 
rise to a specific way of organising things and making decisions (Alessandro 
Narduzzo and Alessandro Rossi, "Modularity in action: GNU/Linux and free/ 
open source software development model unleashed", 2003) [176]. 


Some people claim that the way free software projects are organised is similar to a sur- 
gical team, as proposed by Harlan Mills (of IBM) in the early seventies popularised by 
Brooks in his famous book The mythical man-month (Frederick P. Brooks Jr., 1975) [150], 
Although there may be cases where the development team of a particular free software 
application consists of a designer/developer (the surgeon) and many co-developers who 
perform auxiliary tasks (systems administration, maintenance, specialised tasks, docu- 
mentation) there is never such a strict and defined separation as the one suggested by 
Mills and Brooks. All in all, as Brooks points out in the case of the surgical team, in free 
software the number of developers that need to communicate in order to create a big 
and complex system - the most active ones - is much lower than the total number of 

In the case of the Apache Foundation, we have a meritocracy, since this insti- 
tution has a directors' committee consisting of people who have contributed 
in a notable way to the project. In reality, it is not a strict meritocracy in the 
sense of those who most contribute govern, since the directors' committee is 
elected democratically and regularly by the Foundation's members (responsi- 
ble for managing various free software projects, like Apache, Jakarta, etc.). To 
become a member of the Apache Foundation, you need to have contributed 
in an important and continuous way to one or several of the Foundation's 
projects. This system is also employed by other large projects, such as FreeBSD 
or GNOME. 

Another interesting case of formal organisation is the GCC Steering Commit- 
tee. It was created in 1998 to avoid anyone obtaining control over the GCC 
project (GNU Compiler Collection, GNU's compiler system) and backed by 
the FSF (promoter of the GNU project) a few months later. In a certain sense, 
this committee continues the tradition of a similar one that the EGCS project 
had (which for a time ran in parallel to the GCC project, but later joined it). 
Its fundamental mission is to ensure that the GCC project fulfils the project's 
mission statement. The committee's members are members in a private capaci- 
ty, and are selected by the project itself in such a way as to faithfully represent 

GNUFDL» PID_001 48386 104 Free Software 

the different communities that collaborate in the GCC's development (sup- 
port developers for several programming languages, developers related to the 
kernel, groups interested in embedded programming, etc.). 

The same person does not have to be the leader of a free software project 
forever. Basically, there can be two circumstances in which the project leader 
stops being so. The first is lack of interest, time or motivation to continue. In 
this case, the baton must be passed to another developer who will assume the 
role of project leader. Recent studies (Jesus M. Gonzalez Barahona and Gre- 
gorio Robles, 2003) [87] show that, in general, project leadership frequently 
changes hands, in such a way that we can see several generations of developers 
over time. The second case is more problematic: it involves a forking. Free 
software licences allow code to be taken, modified and redistributed by any- 
body without requiring the project leader's approval. This does not normally 
tend to happen, except in cases where the idea is to deliberately avoid the 
project leader (and the leader's potential veto against a contribution). This is 
similar on the one hand to a sort of "coup d'etat", which on the other hand is 
totally licit and legitimate. For this reason, one of a project leader's objectives 
in keeping co-developers satisfied is to minimise the possibility of a forking. 

7.4. Free software processes 

Although free software is not necessarily associated with a specific software de- 
velopment process, there is a broad consensus about the processes that it most 
commonly uses. This does not mean that no free software projects have been 
created using classical processes, such as the waterfall model. In general, the 
development model of free software projects tends to be more informal, due 
mostly to the fact that a large share of the development team performs these 
tasks voluntarily and in exchange for no financial reward, at least directly. 

The way of capturing requirements in the world of free software depends as 
much on the "age" as on the size of the project. In the early stages, the project's 
founder and the user tend to be the same person. Later on, and if the project 
expands, the capture of requirements tends to take place through electronic 
mailing lists and a clear distinction tends to be reached between the devel- 
opment team, or at least, the more active developers and the users. For large 
projects, with many users and many developers, requirements are captured 
using the same tool as the one used for managing the project's bugs. In this 
case, instead of dealing with bugs, they refer to activities, although the mech- 
anism used for managing them is identical to the one for debugging (they will 
be classified in order of importance, dependency, etc., and it will be possible 
to monitor whether they have been resolved or not). The use of this planning 
tool is fairly recent, so we can see how the world of free software has evolved 
somewhat from a total lack, to a centralised system for managing these activ- 

GNUFDL» PID_001 48386 105 Free Software 

ities in engineering terms, even if it is certainly more limited. All in all, it is 
not usual to find a document that gathers the requirements, as is normally 
the case in the waterfall model. 

As for the system's global design, only large projects tend to have it document- 
ed in comprehensive detail. For the rest, the main developer (or group of main 
developers) is most likely the only one to have it, in their head; sometimes, 
this is even not the case, and the system takes shape as the software evolves. 
The lack of a detailed design not only imposes limitations regarding the pos- 
sible reuse of modules, but also is a large obstacle when it comes to giving 
new developers access, since they will have to face a costly and slow learning 
process. Having a detailed design is not very common either. The lack of it 
means that many opportunities for reusing code are lost. 

Implementation is the phase where free software developers concentrate most 
effort, among other reasons because in their view it is clearly the most fun. 
To do this, the classical programming model of trial and error is normally ob- 
served until the desired results are achieved from the programmer's subjective 
point of view. Historically, it is rare that unit tests are included with the code, 
even if they would make modification or inclusion of subsequent code by oth- 
er developers easier. In the case of certain large projects, such as Mozilla for ex- 
ample, there are machines exclusively dedicated to downloading repositories 
containing the most recent code and to compile it for different architectures 
("An overview of the software engineering process and tools in the Mozilla 
project", 2002) [193]. Detected bugs are notified to a mailing list of developers. 

However, automatic tests are not an entrenched practice. In general, users 
themselves, with their enormous range of uses, architectures and combina- 
tions, will carry them out. This has the advantage of running them in parallel 
at a minimum cost for the development team. The problem with this model 
is how to obtain feedback from users and organise it as efficiently as possible. 

As far as software maintenance in the world of free software is concerned, 
understood as the maintenance of previous versions, having this task will de- 
pend on the project. For projects that need stability, such as operating system 
kernels, previous versions are maintained, since changing to a new version 
can be traumatic. But in general, for most free software projects, if a bug is 
found in a previous version, developers will usually ignore it and recommend 
the use of the latest version in the hope that the bug has disappeared with 
the software's evolution. 

7.5. Criticism of "The cathedral and the bazaar" 

"The cathedral and the bazaar" suffers from not being systematic and a lack of 
rigour given its journalistic rather than scientific nature. The most frequent 
criticisms refer to the fact that it basically explains the particular case of the 
Linux experience and aims to extend those conclusions to all free software 

GNUFDL» PID_001 48386 106 Free Software 

projects. In this sense, in "Cave or community? An empirical examination 
of 100 mature open source projects" [160] we can see that the existence of a 
community as large as the community of the Linux kernel is an exception 
rather than the rule. 

Even more critical are those who believe that Linux is an example of the cathe- 
dral development model. They argue that obviously there is a driving force, 
or at least a person with maximum authority, and a hierarchical system that 
delegates responsibility down to the labourers/programmers. Also, there is a 
distribution of tasks, albeit implicitly. "A second look at the cathedral and the 
bazaar" [91] goes beyond and maintains, not without a certain level of bitter- 
ness and arrogance in its reasoning, that the metaphor of the bazaar is inter- 
nally contradictory. 

Another of the most criticised points of "The cathedral and the bazaar" is its 
assertion that the Brooks law, which states that "adding developers to a de- 
layed software project delays it even more" (The mythical man-month. Essays 
on software engineering, 1975) [150], is not valid in the world of free software. 
In [148] we can read how what happens in reality is that the environmental 
contexts are different and that what in principle appears to be incongruent 
with Brooks' law, after a more comprehensive analysis, is just a mirage. 

7.6. Quantitative studies 

Free software makes it possible to go deeper into the study of code and other 
parameters that intervene in its generation thanks to having access to many 
public information sources. This allows areas of traditional software engineer- 
ing such as empirical software engineering to be fostered due to the existence 
of a huge amount of information that can be accessed without the need to 
heavily intrude in the development of free software. The authors are con- 
vinced that this vision can contribute enormously to the analysis and com- 
prehension of the phenomena associated with free software (and software in 
general), and that it may even, among other possibilities, manage to produce 
predictive software models with feedback in real time. 

The idea behind it is very simple: "given that we have the opportunity to study 
an immense number of free software programs, let's do so." And in addition to 
a project's present status, its past evolution is public, meaning that all of this 
information, duly extracted, analysed and packaged, can serve as a knowledge 
base that allows us to evaluate a project's state of health, helping towards de- 
cision-making and foreseeing current and future complications. 

The first quantitative study of any importance in the world of free soft- 
ware dates back to 1998, although it was published in early 2000 ("The Or- 
biten free software survey") [127]. Its purpose was to find out in empirical 
terms the participation of developers in free software. To do so they sta- 
tistically processed the authorship assignments that authors tend to place 

GNUFDL» PID_001 48386 107 Free Software 

in the heading of source code files. The results showed that participation 
was consistent with the Pareto law ("Course of Political Economy", Lausana, 
1896) [182]: 80% of the code corresponds to the most active 20% of devel- 
opers, whereas the remaining 80% of developers contribute 20% of the to- 
tal code. Many subsequent studies have confirmed and extended the valid- 
ity of this result to different ways of participating in the contribution of 
source code (mailing lists, bug notifications or even the number of downloads, 
as we can see in 
Sourceforge%20paper.pdf [145]). 


The fact that many economic terms appear in the study of free software engineering is 
a result of the interest some economists have shown in learning about and understand- 
ing what motivates volunteers to produce high value goods without usually obtaining 
a direct benefit in exchange. The most well-known article is "Cooking pot markets: an 
economic model for the trade in free goods and services on the Internet" [125], which 
introduces the idea of the gift economy on Internet. At 
Pareto [232] we can obtain further details on the Pareto principle and its generalisation to 
the Pareto distribution. The Lorenz curve ( 
[231], which graphically shows developers' participation in a project, is also interesting as 
well as the Gini coefficient ( [230], cal- 
culated on the basis of the Lorenz curve and which produces a number that shows the 
system's inequality. 

The tool used to conduct this study was published by its authors under a free 
licence, meaning that its results can be reproduced and it can be used to con- 
duct new studies. 

In a later study, Koch ("Results from software engineering research into open 
source development projects using public data", 2000) [158] went further 
and also analysed the interactions in a free software project. The informa- 
tion sources were mailing lists and the repository of versions of the GNOME 
project. But the most interesting aspect of the Koch study was the economic 
analysis. Koch focuses on checking the validity of classical cost forecasts (func- 
tion points, COCOMO model...) and shows the problems involved in applying 
them, although it does admit that the results obtained have to be taken with 
due reserve do partly match reality. He concludes that free software requires 
its own models and methods of study, since known ones are not adapted to 
its nature. However, obviously being able to obtain much of the data related 
to the development of free software publicly, allows us to be optimistic about 
achieving these objectives in the near future. Koch's can be considered the 
first full quantitative analysis, although it certainly lacks a clear methodology, 
and especially some ad hoc tools that would have made it possible to verify its 
results and to study other projects. 

In the year 2000, Mockus et al. presented the first study of free software 
projects encompassing a full description of the development process and or- 
ganisational structures, with both qualitative and quantitative evidence ("A 
case study of open source software development: the Apache server") [172]. To 
do so, they used the software changelog and bug reports to quantify aspects of 
developers' participation, core group size, code authorship, productivity, fault 

GNUFDL" PID_001 48386 108 Free Software 

density, and problem-solving intervals. In a way, this study is still a classical 
software engineering study, save for the fact that the data has been integral- 
ly obtained from the semi-automatic inspection of the data that the projects 
offer publicly on the net. As in the case of "Results from software engineer- 
ing research into open source development projects using public data", 2000 
[158], this article did not provide any tool or automatic process that could be 
reused in future by other research teams. 

In "Estimating Linux's size", 2000 [227], and "More than a gigabuck: estimat- 
ing GNU/Linux's" [228] we find a quantitative analysis of the lines of code and 
programming languages used in the Red Hat distribution. Gonzalez Barahona 
et al. have followed these steps in a series of articles on the Debian distribu- 
tion (vid. for example "Anatomy of two GNU/Linux distributions" [88]). All of 
these provide a sort oi X-ray of these GNU/Linux distributions on the basis of 
data provided by a tool that counts a program's source lines of code (SLOC, 
lines of code that are not blank lines or comments). Aside from the spectacu- 
lar result in total lines of code (Debian 3.0 known as Woody, has more than 
one hundred million lines of code), we can see how the number of lines is 
distributed for each programming language. Being able to study the evolution 
of the different Debian versions over time has thrown up some interesting 
results [88]. It is worth noting that in the last five years the average package 
size has remained practically constant, meaning that the natural tendency to 
grow has been neutralised by the inclusion of smaller packages. At the same 
time, we can see how the importance of the C programming language, though 
still predominant, is declining over time, whereas script languages (Python, 
PHP and Perl) and Java are experiencing an explosive growth. The "classical" 
compiled languages (Pascal, Ada, Modula...) are clearly receding. Finally, these 
articles include a section that shows the results obtained if we apply the clas- 
sical COCOMO effort estimate model dating from the early eighties (Software 
Engineering Economics, 1981) [93] and which is used by proprietary software to 
estimate effort, project schedules and costs. 

Although precursors, most of the studies presented in this section are fairly 
limited to the projects under analysis. The methodology employed has been 
adapted to the analysed project, is partly manual and occasionally the au- 
tomated part can be used generally with other free software projects. This 
means that the effort required to study a new project is much greater, since the 
method needs to be readapted and the manual tasks will have to be repeated. 

For this reason, the latest efforts ("Studying the evolution of libre software 

projects using publicly available data", in: Proceedings of the 3 r Workshop on 

Open Source Software Engineering, 25 l International Conference on Software 
Engineering, Portland, USA [196] or "Automating the measurement of open 
source projects", 2003 [124]) focus on creating an analysis infrastructure that 
integrates several tools so that the process can be automated to a maximum. 
There are two fairly obvious reasons for doing this: the first is that once a lot 

GNUFDL" PID_001 48386 109 Free Software 

of time and effort has been invested in creating a tool to analyse a project with 
special emphasis on making it generic, the effort involved in using it for other 
free software projects is minimal. The second is that analysis using a series 
of tools that study programs from different and sometimes complementary 
points of view, at times does not allow us to obtain a broader vision of the 
project. In the Libre Software Engineering Web Site [86] we can follow these 
initiatives in more detail. 

7.7. Future work 

Having described the brief but intense history of software engineering research 
on free software, we can say that it is still taking its first steps. Many important 
aspects are still pending analysis and detailed examination until we can find 
a model that at least partly explains how free software is generated. The issues 
that will need to be tackled in the near future include the classification of free 
software projects, the creation of a methodology based inasmuch as possible 
on automated analysis and the use of acquired knowledge to build models 
that help us to understand how free software develops at the same time as 
facilitating decision-making on the basis of acquired experience. 

Another aspect that should not be overlooked and that is starting to be con- 
sidered now is the validity of classical engineering methods in the field of free 
software across all software engineering intensifications. Hence, for example, 
the laws of software evolution postulated by Lehman ("Metrics and laws of 
software evolution - the nineties view" [165]) at the beginning of the nine- 
teen seventies and updated and expanded in the eighties and nineties appear 
not to be fulfilled unconditionally in the development of some free software 
projects ("Understanding open source software evolution: applying, breaking 
and rethinking the laws of software evolution", 2003 [199]). 

Currently, one of the most serious deficiencies is the lack of a strict classifi- 
cation so that free software projects can be classed into different categories. 
At present, the classification criteria are too broad, and projects with very dis- 
parate organisational, technical or other characteristics are all put into the 
same bag. The argument that Linux, with an extensive community and large 
number of developers, has a different nature and does not behave in the same 
way as a much more limited project in numbers of developers and users, is 
very true. All in all, a more detailed classification would make it possible to 
reuse the experience acquired in other similar projects (in other words, with 
similar characteristics), making it easier to make forecasts, and making it pos- 
sible to foresee risks, etc. 

The second important aspect that free software engineering needs to tackle, 
closely connected to the preceding point and current trends, is the creation of 
a methodology and tools to support it. A clear and concise methodology will 
make it possible to study all projects on an equal footing, discover their current 
status, learn how they have evolved, and of course, classify them. Tools are 

GNUFDL • PID_001 48386 1 1 Free Software 

essential when it comes to dealing with this problem, since once created they 
make it possible to analyse thousands of projects with minimum additional 
effort. One of the objectives of free software engineering is to make it possible 
to study a project in depth on the basis of a limited set of parameters showing 
where information on the project can be found on the Net (the address of 
the software versions repository, the place where the mailing list archives are 
stored, the location of the bug management system, and a minimum survey). 
Project managers would then be just a button away from a complete analysis, 
a sort of clinical analysis that helped to diagnose a project's state of health 
including at the same time indications on areas for improvement. 

Once we have acquired methods, a classification and models, the opportuni- 
ties arising from simulation, and to be more precise, intelligent agents, could 
be enormous. Considering that our starting point is a notoriously complex 
system, it would be interesting to create dynamic models on which the differ- 
ent entities participating in software generation could be modelled. Obvious- 
ly, the more we know about the different elements, the more adapted to real- 
ity our model will be. Although several proposals for free software simulation 
are known, they are fairly simple and incomplete. To some extent, this is due 
to the fact that there is still an enormous lack of knowledge with regards to 
the interactions that take place in the generation of free software. If we man- 
age to correctly package and process projects' information throughout their 
history, the agents could become crucial for knowing what their future evolu- 
tion will be. Although there are many proposals as to how to approach this 
problem, one of the most advanced for now can be found at http://wwwai.wu- [82]. 

7.8. Summary 

In summary, we have tried to show in this chapter that free software engineer- 
ing is still a young and unexplored field. Its first steps are due to journalistic 
essays that proposed, not without a certain lack of scientific rigour, a more 
efficient development model, but gradually progress has been made towards a 
systematic study of free software from an engineering perspective. Currently, 
following several years of reports and quantitative and qualitative analysis of 
free projects, an enormous effort is being made to achieve a global infrastruc- 
ture that makes it possible to classify, analyse and model the project within 
a limited space of time and in a partly automated manner. When analysing 
free software projects stops being so costly in time and effort as it is now, it 
is likely that a new stage in software engineering will begin, with a different 
type of techniques appearing on the scene designed mainly to predict software 
evolution and foresee potential complications. 

GNUFDL» PID_001 48386 111 Free Software 

8. Development environments and technologies 

"The tools we use have a profound (and devious!) influence on our thinking habits, and, 
therefore, on our thinking abilities." 

Edsger W. Dijkstra, "How do we tell truths that might hurt?" 

Down the years free software projects have created their own (also free) tools 
and systems to contribute to the development process. Although each project 
follows its own rules and uses its own set of tools, there are certain practices, 
environments and technologies that can be considered usual in the world of 
free software development. In this chapter we will look at the most common 
ones and discuss their impact on projects' management and evolution. 

8.1. Description of environments, tools and systems 

Before explaining about specific tools, we will define their general character- 
istics and properties according to the task to be performed and the way devel- 
opers are organised. 

Firstly, although it is not necessarily a determining factor, it is common for the 
environment, development tools (and even the target virtual machine, when 
there is one), also to be free. This has not always been the case. For example, the 
GNU project, with the objective of replacing Unix, had to be developed in and 
for proprietary Unix systems until Linux and the free BSDs appeared. Nowa- 
days, especially when free software is developed as part of a business model, 
the tendency is that the target machine can also be a proprietary system, often 
through interposed virtual machines (Java, Python, PHP, etc.). In any case, the 
environment and the virtual machine need to be sufficiently common and 
cheap to bring together enough co-developers having the same tools. 

Secondly, also in order to attract the largest possible number of co-developers, 
the tools need to be simple, well known and capable of functioning on econom- 
ical machines. Perhaps for these reasons the world of free software is fairly 
conservative when it comes to languages, tools and environments. 

In the third place, the free software development model tends to be eminently 
distributed, with many potential collaborators spread all around the world. For 
this reason generally asynchronous collaboration tools are necessary, which 
at the same time allow the development to progress easily, irrespective of the 
amount and rhythm of work of each collaborator, without delaying anyone. 

Finally, it is advisable to provide developers with various different architec- 
tures on which they can compile and test their programs. 

GNUFDL • PID_001 48386 1 1 2 Free Software 

8.2. Associated languages and tools 

Most free software is written in C language, not only because C is the natural 
language of any Unix variant (the usual free software platform), but also be- 
cause it is widespread, both in people's minds and in the machines (GCC is a 
standard compiler installed by default in almost every distribution). Precisely 
for these reasons and for its efficiency, Stallman recommends its use in GNU 
projects ("GNU coding standards") [203]. Other fairly similar languages are 
C++, also supported by default by GCC, and Java, which has certain similarity 
and is popular because it allows developments for virtual machines available 
in a wide range of platforms. Generally, software engineering reasons are not 
taken into account: in SourceForge (vid. section 8.9.1), in 2004, for every one 
hundred and sixty projects in C there was one in Ada, although the latter is 
supposedly a more appropriate language for developing quality programs. At 
the same time, English is the lingua franca of free software developers, despite 
the fact that Esperanto is a much easier language to learn with a much more 
logical structure. Interpreted languages designed for the rapid prototyping of 
normal applications and web services such as Perl, Python and PHP are also 

Just as C is the standard language, make is the standard program building tool, 
given its source code files. A free programmer will normally use the GNU ver- 
sion (GNU make) [36] rather than BSD's incompatible one (Adam de Boor, 
"PMake - a tutorial") [100]. They can be used to specify dependency trees be- 
tween files, and rules for generating dependent files from those that they de- 
pend on. Thus, we can specify that an object file x . o depends on source files 
x . c and x . h and that to build it we need to execute gcc -c x . c. Or that 
our program's executable depends on a collection of objects and is linked in 
a certain way. When we modify source code and then execute make, only the 
affected modules will be recompiled and the final object will be linked again. 
This is a very low level tool, since, for example, it is incapable of finding out 
for itself when a module needs to be recompiled in C, despite the fact that 
it could do so by examining the chains of includes. It is also very powerful, 
because it can combine all the file transformation tools available in order to 
build very complex targets of a multi-language project. But it is very compli- 
cated and very dependent on Unix- type environments. Other supposedly bet- 
ter alternatives, such as jam (Jam Product Information) [41], aap (Aap Project) 
[1] or ant (The Apache Ant Project) [7] are rarely used (the latter is gaining 
popularity especially in the world of Java). 

Given the heterogeneity of existing systems even in the world of Unix, 
we also use tools designed to help make our programs portable. The GNU 
tools autoconf ( [10], automake (http:/ 
/www. [32] and libtool ( 
ware/libtool) [35] make these tasks easier in C and Unix environments. Giv- 
en the diversity of languages, character sets and cultural contexts, C program- 
mers (and programmers using many other languages) often use gettext (http:/ 

GNUFDL • PID_001 48386 1 1 3 Free Software 

/www. [31] and the internationalisation options of 
the standard C library ( [34] for program- 
ming applications that can be easily localised to any cultural environment at 

Thus, when we receive a source package, it is most likely written in C, packaged 
with tar, compressed with gzip, made portable with autoconf and associated 
tools, and can be built and installed with make. Its installation will be carried 
out in a very similar process to the following one: 

tar xzvf package-1 . 3 . 5 . tar . gz 

cd package-1 . 3 . 5 

. /configure 

make make install 

8.3. Integrated development environments 

An IDE (integrated development environment) is a system that makes software 
developer's work easier by solidly integrating the language oriented edition, 
the compilation or interpretation, debugging, performance measurements, in- 
corporation of source code to a source control system, etc., normally in a mod- 
ular fashion. 

Not all free software developers like these tools, although their use has grad- 
ually expanded. In the world of free software, the first one to be extensively 
used was GNU Emacs ( [33], star work 
of Richard Stallman, written and extensible in Emacs Lisp, for which there are 
lots of contributions. 

Eclipse (Eclipse - An Open Development Platform) [23] can be considered 
today's reference IDE in the world of free software, with the disadvantage 
that it works better (around May 2007) on a non-free virtual Java machine 
(Sun's which is hoped to become free soon anyway). Other popular environ- 
ments are Kdevelop ( [42] for KDE, Anjuta (http:// [6] for GNOME, Netbeans ( [51] of 
Sun for Java and Code::Blocks ( [18] for C++ ap- 

8.4. Basic collaboration mechanisms 

Free software is a phenomenon made possible by the collaboration of dis- 
tributed communities and that, therefore, requires tools to make that collab- 
oration effective. Although for a long time magnetic tapes were physically 
posted, the speedy development of free software began once it became possi- 
ble to communicate rapidly with many people and to distribute program code 
to them or reply with comments and patches. For convenience, rather than 

GNUFDL' PID 00148386 


Free Software 

sending code, messages could be used to send information on the site from 
which the code could be collected. In fact, right in the beginning of the sev- 
enties, e-mail was an extension of the ARPANET file transfer protocol. 

In the world of Unix, in the mid-seventies, uucp, the Unix file transfer pro- 
tocol, was developed for communicating machines through dial-up and ded- 
icated lines, and on which electronic mail was built, and in 1979, the first 
USENET link over UUCP. USENET news, a hierarchically structured forum sys- 
tem distributed by flooding to hierarchically arranged sites, played a funda- 
mental role in the development of free software, sending the source code of 
complete programs to the comp . sources groups. 

Simultaneously, mailing lists were developed, among which the BITNET 
(1981) mailing list managers deserve mention. Nowadays the tendency is to 
prefer mailing lists over USENET-type newsgroups. The main reason has been 
the abuse for commercial purposes and intrusion of "absentminded" people, 
interfering with noise in the discussions. Also, mailing lists provide more con- 
trol and can reach more people. Recipients need to subscribe and any e-mail 
address is valid, even if there is no direct Internet access. The mailing list ad- 
ministrator can choose to know who subscribes or to unsubscribe someone. 
The contributions can be restricted to members only or the programmer may 
choose to moderate the articles before they appear 6 . 

( T"here are also moderated news- 

Traditionally, mailing list administration has been done by e-mail, using spe- 
cial messages with a password, allowing the administrator not to have per- 
manent Internet access, although this is becoming an increasingly rare phe- 
nomenon, meaning that the most popular mailing lists manager nowadays 
(Mailman, the GNU Mailing List Manager) [46] cannot be administrated by 
e-mail, but rather necessarily via the web. The mailing lists play a crucial role 
in the world of free software and in many cases 7 they may be the only way 
to contribute. 

Currently, with the web's popularity, many forums are pure web forums or 
weblogs, with no other interface than the one provided by the navigator. They 
can be generic, like the popular SlashDot (Slashdot: News for Nerds") [58] or 
the Spanish Barrapunto ( [11], where new free software 
is announced or discussed. Or they can be specialised in a specific program; 
in this case they are often integrated with several additional tools in collab- 
oration sites (see section 8.6.2). There are also web interfaces to newsgroups 
and traditional lists. 

( Tor example, contributions 
to Linux have to be made as 
text patches to the list linux- 
kernel@vger . kernel . org . 

Another collaboration mechanism that has become popular at the same time, 
is based on wikis, especially when the idea is to build a joint document, such 
as the specification for a program, a module or a system. We discuss this in 
section 8.6.2. 

GNUFDL • PID_001 48386 1 1 5 Free Software 

Finally, we should mention the interaction mechanisms used by developers 
to converse in real time. For free software it does not tend to be a practical 
mechanism, because with all the developers distributed around the world it is 
not easy to find a convenient time for everyone. Nonetheless, there are sev- 
eral projects that use these text chat tools, either regularly or at virtual con- 
ferences on set dates. The most commonly used tool is the IRC (Internet Re- 
lay Chat, [151], which normally commu- 
nicates people through themed "channels" established on the basis of a series 
of collaborating servers. It is not common for multimedia tools to be used 
(sound, image.,) probably because quality connections are required which not 
everyone may have and that can entail problems with the free software avail- 
able, and the difficulty of registering and editing the results of conversations 
for documenting purposes. 

8.5. Source management 

It is advisable for any program development project to archive its history, be- 
cause a modification could produce a hidden error discovered later for ex- 
ample, and the original needs to be recovered, at least in order to analyse 
the problem. If the project is developed by several people, the author of each 
change will also need to be recorded, for the same reasons as explained above. 
If versioned releases of a project are made, we need to know exactly which ver- 
sions of each module form part of each release. Often, a project will keep one 
stable version and another experimental version; both need to be maintained, 
debugged, and corrected errors transferred from one version to the other. This 
can all be done by saving and labelling each and every version of the files 
correctly, which has generally been considered an excessive cost, although 
with current drives this is becoming less true. What a source control system, 
also known as a version management system, normally does, is to save the file 
history as a set of differences against a version, normally the most recent one, 
for efficiency, also labelling each difference with the necessary metadata. 

But we also want a system of these characteristics to serve for many program- 
mers to collaborate effectively without stepping on each other's toes, but with- 
out impeding each other's progress. Therefore, we need to be able to allow 
several programmers to work concurrently, but with a certain level of control. 
This control can be optimistic or pessimistic. With pessimistic control, a pro- 
grammer can reserve some files to himself to improve for a time, during which 
nobody else can touch those files. This is very safe, but will block other pro- 
grammers and may delay the project, especially if the programmer that has 
locked the files is busy with other things or has even forgotten about them. 
Allowing others to progress is more dynamic, but more dangerous, since in- 
compatible modifications can occur. An optimistic system allows progress to 
be made, but warns us when there have been conflicts and gives us tools to 
resolve them. 

GNUFDL • PID_001 48386 1 1 6 Free Software 

8.5.1. CVS 

CVS (Concurrent Version System) is an optimistic source management system 
designed towards the end of the eighties and used by the vast majority of free 
projects (Concurrent Version System [20], Open source code development with 

CVS, 2 nd edition [113],Version Management with CVS [95]). It uses a central 
repository accessed through a client/server system. The site administrator de- 
cides who has access to the repository, or to which parts of the repository, 
although normally, once a developer has been admitted within the circle of 
trust, he will have acess to all files. Anonymous access, in read-only mode, 
may also be allowed for anyone. 

The anonymous collaborator 

The anonymous CVS is a vital tool for fulfilling the "release early and often" 
concept advocated by Eric Raymond. Any user anxious to try the latest version 
of a program can extract it from the CVS, discover bugs and report them, even 
in the form of patches with the correction. And it can examine the full history 
of the development. 

Let's look a bit at the mechanics. An advanced user wishes to obtain the lat- 
est version of the module mod from an anonymously accessible repository in 
progs .org, directory /var/lib/cvs and protocol pserver. The first time 
he will declare his intention to enter: 

cvs -d :pserver : anonymous@progs . org : /var/lib/cvs login 

If a password is requested, it will be anonymous user (usually the carriage 
return), which will be registered in a local file (this operation is not really 
necessary for anonymous access, but the program will complain if the file with 
the password does not exist). Next, the important thing is obtain the first copy 
of the module: 

cvs -d :pserver : anonymous@progs . org : /var/lib/cvs co mod 

This will create a directory mod with all of the module's files and directories 
and some metadata (contents in subdirectories called cvs), which will allow, 
among other things, not having to repeat the information already provided. 
Our advanced user will enter the created directory, generate the package and 
test it: 

cd mod 

. /configure 


make install . . . 

GNUFDL' PID 00148386 


Free Software 

When he wishes to obtain a new version, he will simply update his copy with- 
in mod. 

cd mod 

cvs update 

. /configure 


make install . . . 

If he finds a bug, he can correct it in place and then send a patch via e-mail 
to the program's maintainer (individual or mailing list): 

cvs diff -ubB | mail -s "My patches" 

The normal developer 

The normal developer will have an account on the server. He can use the 
same mechanism and the same protocol as the anonymous user, replacing 
anonymous for his account name. 

Once he has a working copy of the module, he can make the necessary 
changes, and when he considers that they have been stabilised, commit the 
changes to the repository. For example, if he modifies the files part . h and 
part . c, he will commit them like this: 


For security reasons, for ac- 
counts with write permissions, 
ssh tends to be used, as it 
provides an authenticated and 
encrypted channel. 

cvs ci part.h part.c 

Before completing the operation, the CVS will ask him for an explanation of 
what he has done, which will be attached to both files' log. Also the revision 
number of each file will be increased by one unit. This number identifies every 
important moment in the history of a file and can be used to recover each 
one of those moments. 

When should a developer do a commit? This is a question of methodology 
that project members need to agree, but it seems obvious that changes that do 
not compile should not be committed. But it is preferable to pass also a min- 
imum test battery. In many projects the approval of a project or sub-project 
supervisor who examines the modification is also required. 

In developing the modification, someone may have altered other files, or even 
the same ones. Therefore it is advisable for developers to do a relatively fre- 
quent update of their copy (cvs update). If other files have been modified, the 
environment may also have changed and tests that were previously passed 
may now be failed. If the same files have been modified, it could be that these 
changes have occurred either in places or routines that we have not touched 
or in code that we have modified. In the first case there is no conflict (at 

GNUFDL • PID_001 48386 1 1 8 Free Software 

least not apparent) and the modification operation "merges" our version with 
the repository's, generating combined files, with all of the changes. Otherwise 
there is a conflict, in which case we need to discuss with the developer who 
has made the other changes and agree to a final version. 

For better identification of each project component, it is advisable for it to < 8 >i n cvs the revision numbers nor- 

carry directly associated revision information. CVS can label source codes and mally ^ ave tw ° components (ma- 
1 J jor and minor), but they can have 

objects automatically, on condition of following a certain discipline. For ex- four, six, etc. 
ample, if in a source code comment we write the key word $id$, every time 
the file is committed to the repository, the word will be replaced with an iden- 
tification chain that will show the file name, the revision number 8 , the date 
and time of the commit and its author: 

$Id: part.c,v 1.7 2003/07/11 08:20:47 Joaquin Exp $ 

If we include this keyword in a string of the program, when compiled the 
string will appear in the object and in the executable, making it possible to 
identify it with a tool (ident). 

The administrator 

Obviously, administrators are responsible of the most complicated part of 
maintaining the repository. For example, they need to register the program, is- 
sue permissions for developers and coordinate them, label delivered versions, 

It is common practice for all projects to have a stable version and an exper- 
imental version. To do this we create branches. Whereas those dedicated to 
maintenance correct errors on the stable branch, new developments are made 
on the experimental branch. When the experimental branch stabilises, it is 
passed onto stable, but not without previously applying the corrections made 
to the former stable branch. This operation is called merging, it is delicate and 
supported by CVS, although in a somewhat primitive way. This idea can be 
extended to the concept of experimental branches which evolve in different 
directions, which may or may not come to a good end, and that in any case, 
unless they are dead ends, will have to be fully or partly integrated into the 
stable product, with appropriate merges. 

A right that free software gives us is to modify a program for private use. Al- 
though it is desirable to contribute all improvements to the common pool, 
often the modifications we wish to make are too specific and uninteresting 
for the public at large. But we are interested in incorporating the evolution in 
the original program. This can be done with a special type of branching and 
merging (vendor branches). 

GNUFDL' PID 00148386 


Free Software 

The administrator can also facilitate team coordination through automated 
mechanisms, such as by generating e-mail messages when certain events oc- 
cur, like commits, or forcing certain automatic actions to be carried out before 
a commit, such as automatic checks of style, compilations, or tests. 

8.5.2. Other source management systems 

Despite being the most extensively used version control system, CVS has some 
notable disadvantages: 

1) CVS does not support either renamings or file directory changes, or meta- 
data (owner, permissions, etc.) or symbolic links. 

2) Because it is an evolution of a version control system for individual files, 
it naturally does not support version control for complete groups. 

3) CVS does not support sets of coherent changes. Indeed, adding a feature 
or correcting an error can involve changing several files. These changes 
should be atomic. 

4) In CVS the use of branches and merges is fairly complicated. In fact, if 
we create an experimental branch of a project and wish to include the 
corrections made to the stable version, we need to know in detail which 
corrections have been made already and which not, so as not to do them 
several times over. 

5) CVS depends on a centralised server, and although it is possible to work 
without a connection, we do need one for generating versions, comparing 
and merging them. 

6) CVS does not generate, without the help of separate tools, the file 
changelog, which shows the global history of a project's changes. 

7) CVS does not support well projects with a very large number of files, as 
in the case of the Linux kernel. 

And yet, there are other free systems which solve several of these problems. We 
would highlight the already mentioned successor of CVS, Subversion (http:// [62], ( [96], which strict- 
ly solves the basic problems of CVS and can use HTTP extensions (WebDAV) 
in order to bypass aggressive security policies. 


In 2007 Subversion is al- 
ready the clear successor of 
CVS, and many free software 
developments have migrated 
to it. 

The development model based on a centralised repository, although suitable 
for cooperative work, does not satisfy all expectations, since being able to cre- 
ate our own development branches depends on the one hand on the server's 
accessibility and good functioning and on the other on the administrators of 
that server. Sometimes distributed repositories are required that allow anyone 

GNUFDL' PID 00148386 


Free Software 

to have a repository with a private or public branch that can be merged or not 
with the official one. This is how GNU arch (Arch Revision Control System) [8] 
or bazaar (Bazaar GPL Distributed Version Control Software) [12] work, as well 
as the proprietary system BitKeeper (Bitkeeper Source Management) [14], cho- 
sen by Linus Torvalds to maintain Linux since February 2002, since according 
to him there was no appropriate free tool. It is said that using Bitkeeper dou- 
bled the pace of development of Linux. Nonetheless, the decision came un- 
der heavy criticism because it was proprietary, with a licence that allowed free 
projects to obtain it free of charge on condition that all commit changes with 
their metadata were logged on a public server designated by the owners and 
accessible to everyone, and always on condition that the licensee did not try 
to develop another source control system to compete with it. It was precisely 
the attempt to develop a compatible free product by an employee of the same 
company where Linus Torvalds worked that detonated the change in source 
management system. Linus rapidly developed a provisional replacement, git 
("Git manual page") [218], which soon became definitive, condensing all of 
the experience of Linux's cooperative and decentralised development: it sup- 
ports large-size projects in a decentralised fashion, facilitating to a great extent 
the development of tentative branches and their merging with others or with 
the main one, with cryptographic security mechanisms that prevent altering 
the log. As of April 2005, Linux is maintained using git or its wraps (for exam- 
ple, cogito "Cogito manual page" [90]. 

8.6. Documentation 

In the world of free software, WYSIWYG text processors and other office suite 
tools that are so successful in other environments are barely used, even though 
there are already free tools such as This is due to several im- 
portant factors: 

It is advisable to maintain documentation under source control, and 
source control systems, like CVS, although they admit binary formats, pre- 
fer transparent text formats that can be edited with a normal text editor 
and processed with tools developed for programs that allow us to see the 
differences between versions easily, to generate and apply patches based 
on those differences, and to carry out merges. 


In Unix the most common 
tools for these operations are 
diff, diff3, patch and merge. 

Some free documentation licences, especially the GFDL (vid. section 
10.2.1), demand transparent formats, especially because they make the job 
easier for those who prepare derived documents. 

The WYSIWYG tools ("what you see is what you get") generally do not con- 
tain any information other than the strict visualisation, making it very dif- 
ficult, if not impossible, to identify authors, or titles, or conversion to oth- 
er formats. Even if they do allow conversion to other formats, this tends 

GNUFDL" PID_001 48386 121 Free Software 

to be done interactively, and is often impossible to automate (using make, 
for example). 

• In general, office applications generate sizeable file formats, which is an 
undesirable feature for both developers and repositories. 

For all of the above, free programmers, accustomed to programming and com- 
piling, prefer transparent document formats, in many cases pure simple text 
and in many others processable document formats. 

The processable formats in use are not many. Traditionally, in the world of 
Unix programs have been documented in the formats expected by the family 
of processors roff, with a free version (GNU troff) [37] by Norman Walsh. Nev- 
ertheless, this practice has been gradually abandoned, except for traditional 
manual pages, since it is almost obligatory to prepare manual pages for the 
system's most basic tools. Because many manual pages have grown so much 
so that it is barely appropriate to call them pages, it was necessary to prepare 
an alternative hypertext format that allowed documents structured with in- 
dexes and cross-references to be followed. The GNU project designed the tex- 
info format (Texinfo - The GNU Documentation System) [63] and made it its 
standard. This format allows navigable documents to be obtained with the 
info tool or within the emacs editor, and in turn, to obtain quality document 
printouts using the TeX text processor, of Donald Knuth (The TeXbook) [156]. 

The texinfo format can be translated into multipage HTML if required, and 
many people prefer to view the information with a web navigator, but the 
capacity to search for words in a document is lost. This is one of the unwanted 
results of the popularity of HTML, since the navigators do not implement the 
concept of multipage document , despite the fact that there are link elements 
that allow parts to be interlinked. 

There is overwhelming demand for being able to view complex documents as 
easily navigable multipage web pages. There are people who write documen- 
tation in LaTeX (LaTeX user's guide and reference manual) [163], also a TeX ap- 
plication, very popular among scientists, more expressive than Texinfo and 
convertible to multipage HTML with certain tools (The LaTeX Web Compan- 
ion) [130], on condition a certain discipline is maintained. Indeed, TeX ap- 
plications are sets of macros that combine very low level typographic opera- 
tors to convert them into abstract languages that work with high level con- 
cepts (author, title, summary, chapter, section, etc.). If we only use the basic 
macros, conversion is simple. But since nobody prevents the use of low level 
operators and, additionally, there are enormous quantities of macro packages 
beyond the maintenance capacity of conversion tool maintainers it is difficult 
to achieve good conversions. 

GNUFDL» PID_001 48386 122 Free Software 

8.6.1. DocBook 

The problem stems from the fact that there is no distinction between con- 
tent and presentation, either in TeX or in nroff, since the abstraction is built 
in layers. This distinction is made by SGML applications {standard generalised 
markup language) [81] and XML {extensible markup language) [224], where the 
presentation is specified with completely separate style sheets. Soon very sim- 
ple SGML applications started to be used, such as linuxdoc and debiandoc, 
but due to their limited expressive capacity, DocBook was chosen. {DocBook: 
the definitive guide) [225] . 

DocBook is an SGML application originally developed for technical IT docu- 
mentation and now has an XML variant. Currently, DocBook is the standard 
free documentation format for many projects (Linux Documentation Project, 
KDE, GNOME, Mandriva Linux, etc.) and a goal to be reached for others (Lin- 
ux, *BSD, Debian, etc). 

However, DocBook is a complicated language, plagued by tags, which means 
that it is useful to have tools to help with the editing, even if they are very 
basic and rare; one of the most popular tools of this type is the psgml mode 
of emacs. It is also heavy to process and free processors still generate a not very 
attractive quality of documents. 

8.6.2. Wikis 

Many people find it too complicated to write documentation with such com- 
plex languages as DocBook and collaboration mechanisms like CVS. This is 
why a new mechanism of collaboration for online document preparation via 
the web has become popular, called wiki, and invented by Ward Cunningham 
("Wiki design principles") [97]. It was first put into service in 1995 and is now 
extensively used in a wide range of variants for preparing very dynamic doc- 
uments, not designed for printing and often with a short life (for example, 
conference organisation). 

Unlike DocBook, a wiki has a very simple and concise markup language which 
is reminiscent of the final presentation, without being exactly like it. For ex- 
ample, paragraphs are separated by a blank line, elements of a list are started 
with a hyphen if not numbered and with a zero if they are numbered, and 
table cells are separated by vertical and horizontal bars. 

Neither does the concept of a "full document" exist, rather a wiki is more a set 
of small interlinked documents created as and when it is necessary to explain 
a new concept or subject. The documents are created almost automatically, as 
the editing tool shows very clearly that we have entered a concept (through 
a wikiName, almost always two joined words with the first letter capitalised). 
Hardly any wiki allow hyperlinks within the same page. 

GNUFDL» PID_001 48386 123 Free Software 

Unlike CVS, anyone can write in a wiki, although it is advisable for the author 
to identify himself by previously registering. When we visit a wiki we can see 
that all pages have a button that allows them to be edited. If pressed, the 
navigator will show us a form with the document's source code, which we will 
be able to change. This is not a WYSIWYG edit, which discourages anyone 
just wanting to interfere, but is simple enough for anybody interested to be 
able to modify documents with very little effort. 

Wikis cany their own document version control, in such a way that all of their 
versions are generally accessible, indicating who made them and when. They 
can also be easily compared. Plus, they tend to include search mechanisms, at 
least per page name and word content. 

Normally, the original author of a page will want to know what changes are 
made to it. To do so he can subscribe to the changes and receive notifications 
of them by e-mail. Sometimes, the person seeing a document will not dare 
to change anything, but may make a comment. Normally, all wiki pages have 
an associated comments forum pasted at the end of the document, which 
either the original author or anybody who assumes the role of editor can use 
to reform the original text, possibly by moving phrases from the comments 
to the relevant places. 


The best way of understanding the wiki concept is to access one and experiment on a 
page designed for this purpose, usually called SandBox. 

8.7. Bug management and other issues 

One of the strong points of the free development model is that the commu- 
nity contributes with bug reports and feels that those reports or solutions are 
given attention. This requires a simple bug reporting mechanism, so that de- 
velopers can receive sufficient information, in a systematic way and contain- 
ing all necessary details, either provided by the collaborator, with an explana- 
tion of what is happening, the level of importance and possible solution, or 
through an automatic mechanism that determines, for example, the program 
version and environment in which it functions. Errors should also be saved 
in a database that can be consulted, to see whether a bug has already been 
communicated, corrected, its level of importance, etc. 

There are several of these systems, with different philosophies. Some are via 
web, others via e-mail, through some intermediary program. They all have a 
web interface for consultation. Some allow anonymous reports, while others 
require identification (a valid e-mail address) to prevent noise. Although web 
procedures would appear to be the most simple, they do not easily obtain 
automatic information on the bug's environment. For example, the Debian 
system provides programs like reportbug, which after asking for the name of 
the package that we wish to report on, consults the error server for the bugs 

GNUFDL" PID_001 48386 124 Free Software 

reported to it. If none of them refers to our problem, we will be asked for a de- 
scription of it, its level of importance ("critical", "grave", "serious", "important", 
"cannot be regenerated from source codes", "normal", "minor" or "suggestion") 
and labels about its category (for example, "security"). Following this, if we 
confirm the request, it will automatically find out the version of the package 
and those on which it depends, in addition to the kernel's version and archi- 
tecture. Obviously, it knows the e-mail address, so it sends to the correct site 
a report similar to the following one: 

Package: w3m-ssl 
Version: 0.2.1-4 
Severity: important 

After reloading a page containing complex tables several dozen times, w3m had used 
all physical memory and thrashing commenced. This is an Alpha machine. 

--System Information 

Debian Release: testing/unstable 

Kernel Version: Linux romana 2.2.19 #1 Fri Jun 1 18:20:08 PDT 2001 alpha unknown 

Versions of the packages w3m-ssl depends on: 

ii libc6.1 2.2.3-7 GNU C Library: Shared libraries and Timezone data 

ii libgc5 5.0.alpha4-8 Conservative garbage collector for C 

ii libgpmgl 1.19.3-6 General Purpose Mouse Library [libc6] 

ii libncurses5 5.2.20010318-3 Shared libraries for terminal handling 

ii libsslO.9.6 0.9.6a-3 SSL shared libraries ii w3m 0.2.1-2 WWW browsable pager with 

tables/frames support 

This message generates a bug number which is returned to us, sent to the 
maintainer and saved in the database. When the bug is solved, we will also 
receive a notification. Every bug has an e-mail address assigned to it that can 
be used to provide additional information, for example. We can consult the 
bug database at any time. 

Sometimes bug monitoring systems have mechanisms for assigning someone 
to solve them and setting a deadline. There are also other issues, such as 
pending jobs, requested improvements, translations, etc., that require similar 
management mechanisms. With free software we cannot generally use very 
rigid mechanisms for managing the tasks that each developer has to do. After 
all, many collaborators are volunteers and cannot be obliged to do anything. 
Nonetheless, tasks can be defined and we can wait for somebody to subscribe 
to the system and to take them on within a declared period. Whether there is 
control over what certain people can do or not, it is always advisable to con- 
trol all the tasks that need to be done, who and what they depend on, their 
level of importance, and who is working on them. Many important projects 
manage these aspects using Bugzilla (The Bugzilla guide) [89] or its derivatives. 

GNUFDL» PID_001 48386 125 Free Software 

Sometimes someone working on a project may discover a bug on a different 
project on which his work depends, but that has a different bug management 
system to the one to which he is accustomed. This is particularly true for users 
of distributions who wish to use a single tool for reporting and monitoring 
bug solving. To facilitate reporting and monitoring of those bugs, it may be 
advisable to federate different systems, as done by Malone (The Malone Bug 
Tracker) [47]. 

8.8. Support for other architectures 

The minimum support required for working with a portable program is access 
to compilation farms , which allow the program to be compiled on different 
architectures and operating systems. For example, SourceForge (vid. section 
8.9.1) offered for a time Debian GNU/Linux environments for Intel x86, DEC 
Alpha, PowerPC and SPARC, in addition to Solaris and Mac OS/X environ- 

It is also useful to be able to test (not just compile) the program in those 
environments. But this service requires more resources and more of the 
administrator's time. The compilation service can already be problematic, be- 
cause normally we need to provide compilation environments for several lan- 
guages, with a large number of libraries. If what we want to do is to test any 
program, the difficulties increase exponentially, not just because it is very 
difficult to have the necessary resources available, but also for security rea- 
sons, which can make it extremely complicated to administrate those systems. 
Notwithstanding, there are a few compilation farm services, with standard in- 
stallations of various architectures, which can allow us to test some things. 

The abovementioned public farms are normally a service that requires manual 
use. The invited developer copies his files onto one of those machines, com- 
piles them and tests the result. He will probably have to do it from time to 
time, prior to releasing an important version of the program. It could be much 
more interesting for compilations and the execution of regression tests to be 
carried out systematically, in an automated fashion, for example every night, 
if there have been changes in the source codes. This is how some important 
projects operate, which provide their own infrastructure for external develop- 
ers, which tends to be called a tinderbox. This is the case with Mozilla, financed 
by Netscape, whose tinderbox ( [50] 
has a web interface to the results of the compilation and tests of the navigator's 
components on all of the architectures on which it operates. This interface 
is closely related to the CVS and shows those results for different states (be- 
tween commits), identifying the one responsible for the bugs, and facilitating 
progress, by overcoming the problem until it is resolved. Tinderboxes are also 
used by the projects OpenOffice and FreeBSD, at least. 

GNUFDL» PID_001 48386 126 Free Software 

8.9. Development support sites 

Development support sites offer, in a more or less integrated fashion, all of 
the services described above plus a few additional ones that allow projects to 
be searched by categories and to classify them according to some simple pa- 
rameters of activity This spares the developer having to set up and administer 
an entire infrastructure for collaboration, allowing him to concentrate on the 

8.9.1. SourceForge 

With regards to this type of service, one of the first to become established, 
and the most popular, is SourceForge ( [61], managed 
by the OSDN (Open Software Development Network), a subsidiary of VA Soft- 
ware, which in March 2007 hosted more than 144,000 projects. It is structured 
around a set of programs with the same name, and which up to version 2 were 
free software. 

SourceForge, as a prototype for this type of sites, offers a web interface or glob- 
al access portal ( and a subportal per project (http:// The global interface shows news, advertisements, 
links, and an invitation to become a member or to enter if we already are 
members. To collaborate on the site, it is advisable to become a member, and 
it is compulsory if we want to create a new project or to participate in an ex- 
isting one. To be a spectator it is not necessary, and as such, we can see what 
are the projects experiencing most active development or downloaded most 
frequently, and search for projects by category or descriptive word, and they 
will appear in order of activity level. For each project we can see its descrip- 
tion, status (alpha, beta, production), its descriptors (programming language, 
operating system, subject, type of users, language, licence...), bugs and pend- 
ing or reinstated aspects, activity levels over time..., or download it. We can 
also take part in forums or report bugs, even anonymously, which is not very 
advisable (because, for example, we may not get a reply). 

Any authenticated user can request to register a project, which the adminis- 
trators will admit on condition that it fulfils the site's policies, which in the 
case of SourceForge are fairly liberal. Once authorised, the creator can register 
other users as additional administrators or as developers, with access to mod- 
ify the sources. Following authentication, there are not many more controls 
over the project, which means that there are a lot of dead projects. This does 
not confuse users too much though, because project searches sort the projects 
by level of activity, meaning that low or nil activity projects are barely visible. 
These projects run the risk of being eliminated by the site owners. The services 
that SourceForge offers a project, and that we could expect from any other 
similar service are as follows: 

GNUFDL» PID_001 48386 127 Free Software 

• Hosting for the portal web pages of the project, at the address, for viewing by the public. These pages can be stat- 
ic or dynamic (with CGI or PHP), in which case they can use a database 
(MySQL). They are entered directly through remote copy commands and 
can be handled using remote terminal interactive sessions (SSH). 

• Optionally, a virtual server that responds to addresses from a separately 
obtained domain, like or 

• As many web forums and/or mailing lists as necessary in the 
administrator's opinion. 

• A news service where administrators announce advances concerning the 

• Trackers for bug reporting and monitoring, requests for support, requests 
for improvements or integration of patches. Administrators give the issue 
a priority level and assign a developer to find the solution. 

• Task managers, similar to trackers, that allow sub-projects to be defined 
with a series of tasks. These tasks, in addition to a priority level, are given 
a deadline. From time to time, developers assigned these tasks can show 
percentages of task completion. 

• A CVS or Subversion with initial access rights for all developers. 

• Uploading and downloading service for software packages. It registers en- 
tered versions when used and interested parties can receive a notification 
when this occurs. Plus, the initial upload involves the creation of several 
replicas worldwide, which facilitates distribution. 

• Service for publishing documents in HTML format. Anyone can register 
them, but they will only be visible following approval by an administrator. 

• Back-up copy for disaster recovery, such as broken drive, not user bugs, 
like accidentally deleting a file. 

• Integrated mechanism for donations to users, to projects and to Source- 

An authenticated user will have a personal page containing all relevant infor- 
mation, such as projects to which the user is associated, themes or tasks pend- 
ing, as well as forums and files that he has said he wants to supervise. Plus, so 
that he does not have to be tending to his personal page, the user will receive 
notifications to his e-mail about the things he wishes to control. 

GNUFDL» PID_001 48386 128 Free Software 

8.9.2. SourceForge heirs 

In 2001, VA Software was about to go bankrupt, in the full swing of the dot- 
com crisis. Then it announced a new version of its SourceForge software with 
a non-free licence, in an attempt to secure a source of revenue by selling it 
to companies for their internal developments. At the same time, it eliminat- 
ed mechanisms that allowed a project to be dumped for moving to another 
site. Both events were seen as a threat that the thousands of projects host- 
ed by SourceForge would become trapped in the hands of a single company, 
which would use the platform for showing non-free software. In the face of 
this and the possibility of the site closing, offspring of the free version were 
developed and portals based on it were opened, particularly Savannah (http:/ 
/savannah. gnu. org) [57], dedicated to the GNU project and to other programs 
with copyleft-type licences, or BerliOS (BerliOS: The Open Source Mediator) 
[13], conceived as a meeting point for free software developers and compa- 
nies. However, this is just a step in the direction of developing a distributed 
and replicated platform, where nobody has absolute control over the projects 
(Savannah The Next Generation, 2001) [98]. 

Another example of a free software project management system is Launchpad 
( [43], used by Ubuntu for developing each version of 
the distribution. Launchpad is not a repository for source code, it is designed 
rather to offer support for monitoring code, incidents and translations. To 
achieve this it uses the already mentioned Malone tool, which allows incidents 
to be redirected to each code repository of the affected modules. 

8.9.3. Other sites and programs 

Naturally, collaboration systems have been and continue to be developed, 
and some companies base their business on maintaining and servicing those 
sites. For example, the Tigris project ( Open Source Software En- 
gineering Tools) [64], which not only maintains free software engineering 
projects, it also uses a collaboration portal (SourceCast) maintained by a ser- 
vice company (CollabNet), which also maintains individual projects' sites, like Emerging new sites adopt new free software, such as GForce 
( [30], used by the Debian project ( 
[5]. We can see a detailed comparison of many sites in "Comparison of free/ 
open source hosting (FOSPhost) sites available for hosting projects externally 
from project owners" [202]. 

GNUFDL" PID_001 48386 129 Free Software 

9. Case studies 

"GNU, which stands for 'Gnu's Not Unix', is the name for the complete Unix-compatible 
software system which I am writing so that I can give it away free to everyone who can 
use it. Several other volunteers are helping me. Contributions of time, money, programs 
and equipment are greatly needed." 

Richard Stallman, "The GNU Manifesto" (1985) 

This chapter provides a more in-depth study of some of the most interest- 
ing free software projects in terms of the impact on the free software world, 
the results obtained, the management models, historical development, etc. Of 
course, the number of projects that we can discuss here is much smaller than 
the total number of free software projects (dozens of thousands), which means 
that this chapter should not be thought of as comprehensive, and neither can 
it ever be. Nevertheless, we hope that readers, having read the chapter, will at 
least have a basic understanding of how the theories that we have discussed 
throughout this book have been put into practice. 

The projects that we have chosen range from lower-level applications, the 
ones which interact more with the computer's physical system rather than the 
user, to work environments designed for the end user. We have also included 
free software projects that, in principle, are not strictly development projects. 
This mainly applies to the distributions, which tend to be used as integrating 
systems, as they mainly take an extensive but limited set of independent ap- 
plications and use them to create a system in which everything interacts ef- 
fectively, including the options for installing, updating and deleting applica- 
tions, as desired by the user. 

The lowest-level projects that we will look at will be Linux, the kernel of 
today's most popular free operating system and FreeBSD, which combines the 
kernel from the BSD family with a series of applications and utilities made by 
third parties. The work environments for end users that we will study will be 
KDE and GNOME, which are certainly the most widely-used and popular. For 
the servers, one of the main aspects in free systems, we will look at Apache, 
the leader in the WWW servers market, in this chapter. Likewise, we will in- 
troduce Mozilla, one of the WWW clients (it is in fact, much more than that) 
that we can rely on in the free software world. The last project that we will 
look at in this chapter is, a free Office IT (suite) package. 

We thought it would be appropriate to study the details of two of the most 
popular distributions, Red Hat Linux and Debian GNU/Linux, and to compare 
their sizes to other widely used systems, such as Microsoft Windows or Solaris. 
Finally, the Eclipse multi-language software development environment has 
also been included. 

GNUFDL" PID_001 48386 130 Free Software 

After discussing the different case studies, we provide a table showing the most 
important characteristics of each application or project. One of the elements 
that readers will probably find most surprising will be the results of the cost 
and duration estimations and the number of developers required. We have 
obtained these results using methods typically used in the field of software 
engineering, especially the COCOMO Software Cost Estimation Model. The 
COCOMO model (Software Engineering Economics, 1981) [93] takes the num- 
ber of source code lines as the starting measurement and generates estimates 
of the total cost, the development time and effort required to create the soft- 
ware. COCOMO is a model designed for "classical" software generation pro- 
cesses (waterfall or V model developments) and for average-size or large-scale 
projects; therefore, the figures that it will produce for some of the cases we 
analyse should be taken with some reservations. In any event, the results can 
help to give us an idea of the sheer scale on which we are working and of 
the amount of strenuous effort that would be necessary to achieve the same 
results with a proprietary software development model. 

In general, it is the cost estimates that are most striking out of all the figures 
resulting from the COCOMO model. Two factors are taken into account in 
this estimate: a developer's average salary and the overheads. For calculating 
the estimated costs, the average salary for a full-time systems programmer is 
taken from the year 2000 "Salary survey 2000" [235]. The overheads are the 
extra costs that all companies must pay so that the product can be released, 
independently of the salary paid to the programmers. This ranges from the 
salaries of the secretaries and the marketing team to the costs of the photo- 
copies, lighting, hardware equipment, etc. To summarise, the cost calculated 
by COCOMO is the total cost that a company would have to incur in order to 
create software of the specified dimensions and it should be remembered that 
only a part of this money would be received by the programmers for designing 
the software. Once this is factored in, the costs no longer seem so excessive. 

9.1. Linux 

The Linux kernel is, without a doubt, the star application of free software, to 
the extent that, whilst only constituting a small part of the system, its name is 
used to define the whole. Furthermore, it could even be said that free software 
itself is confused with Linux on many occasions, which is a pretty big mistake 
to make, given that there is free software that runs on systems not based on 
Linux (in fact, one of the biggest aims of the movement and of many free 
software projects is to create applications that can run in numerous environ- 
ments). On another note, there are also applications that work in Linux and 
that are not actually free software (such as Acrobat Reader, the proprietary PDF 
documents reader, for which there is also a Linux version). 

GNUFDL» PID_001 48386 131 Free Software 


There are actually various projects that integrate and distribute free applications that 
run on Windows systems, to avoid free software becoming associated solely with Lin- 
ux systems. One of the pioneers in this area (and the one that probably became most 
well-known and comprehensive) was GNUWin, which was distributed on self-bootable 
CDs with more than a hundred free applications for Win32 systems. Most of these appli- 
cations are also available in common GNU/Linux distributions, which made GNUWin 
a good tool for preparing for a gradual and easy transition from a Windows system to 
a GNU/Linux one. As at early 2007, there are other similar systems available, such as 

9.1.1. A history of Linux 

The history of Linux is one of the most well-known histories within the world 
of free software, most probably because it has the traits of a legend rather than 
those of the history of a computer programme. In 1991, a Finish student called 
Linux Torvalds decided that he wanted to learn how to use protected mode 
386 on a machine that his limited income had allowed him to purchase. At 
that time, there was a kernel in the operating system called Minix, designed for 
academic purposes and for use in university courses on operating systems; this 
is still used today. Andrew Tanenbaum, one of the most prestigious professors 
at the university, was the leader of the team working on the development of 
Minix, based on traditional Unix systems. Minix was a limited system, but 
quite capable and well-designed, and was at the centre of a large academic and 
engineering community. 

Minix had a free distribution license and could be used for academic purposes, 
but it had the big disadvantage that people that did not work or study in the 
University of Amsterdam could not add improvements to it; instead these im- 
provements had to be made independently, usually using patches. This meant 
that in practice, there was an official version of Minix that everybody used 
and then a long series of patches that had to be applied later to obtain addi- 
tional functions. 

In mid- 1991, Linus, then an anonymous Finnish student, sent a message to 
the Minix newsgroup announcing that he was going to start work on an op- 
erating system kernel based on Minix, from scratch, rewriting code. At the 
time, although Linus did not explicitly say that he was going to publish it 
with a free software license, he noted that the system that he was going to 
create would not have the barriers that Minix had; this would indicate that, 
unbeknown to him, and probably without actually wanting to, he was taking 
the first step towards making the community that congregated around Minix 
at that time his. 

Version 0.02, which dates from October 1991, despite being very limited, 
could already execute bash terminals and the GCC compiler. Over the course 
of the following months, the number of external contributions grew to the 
point that in March 1992, Linus could publish version 0.95, which was widely 
acknowledged as almost stable. There was still quite a way to go, however, be- 
fore version 1.0, which is usually considered the first stable one. In December 

GNUFDL' PID 00148386 


Free Software 

1993, for example, version 0.99pll4 was published (which would make it the 
fourteenth corrected version of version 0.99); in March 1994, Linux 1.0 was 
finally born. By this time, Linux was being published under the terms of the 
GPL license; according to Torvalds himself, this was one of the best decisions 
he ever made, as it was extremely helpful in distributing and popularising his 
kernel. In "Evolution in open source software: a case study", [128] there is an 
exhaustive analysis of the evolution of the different versions of the Linux ker- 
nel, focusing on the scale and modularity. 


Another significant event in the annals of free software was the debate that took place 
in late January 1992 on the Minix newsgroup between Andrew Tanenbaum and Linus 
Torvalds. Tanenbaum, who was probably a bit annoyed by Torvalds' success with his 
"toy", attacked Linux and Linus in a rather disproportionate manner. His essential point 
was that Linux was a monolithic system (the kernel integrates all the handlers and the 
rest) and not a microkernel system (the kernel has a modular design, which means that 
it can be much smaller and that modules can be loaded upon demand). The original 
argument can be read just as it occurred in "The Tanenbaum-Torvalds debate" newsgroup 

9.1.2. Linux's way of working 

The way Torvalds worked was not very common at that time. The develop- 
ment was mainly based on a mailing list 9 . The mailing list was a place where 
people not only argued, but where developments also took place. And this was 
because Torvalds was extremely keen on having the whole life of the project 
reflected on the mailing list, which is why he would ask people to send their 
patches to the list. Contrary to what one might have expected (the patches 
sent as attachments), Linus preferred to have the code sent in the body of 
the message so that he and others could comment on the code. In any case, 
although many people would provide their opinions and send corrections or 
new functions, the last word would always go to Linus Torvalds, who would 
decide on what code would be incorporated into Linux. To a large extent, this 
is still how it works in 2007. 

( T"he list's email is linux- The histor- 
ical messages can be seen at http:/ 
/www. uwsg. 


The consolidation of Linus Torvalds as a "benevolent dictator" has given rise to a large 
number of anecdotes within the project. For example, it is said that if an idea is liked, 
it must be implemented. If it is not liked, it must also be implemented. The corollary, 
therefore, is that good ideas are of no use whatsoever (without code, of course). On an- 
other note, if the implementation is not well-liked, it is essential to insist. A well-known 
case is that of Gooch, for whom Saint Job was a mere learner. Gooch made up to one 
hundred and forty six parallel patches until Linus finally decided to integrate them into 
the kernel's official branch. 

Another one of Torvalds' innovative ideas was to develop two branches of 
the kernel in parallel: the stable one (the second number of the version is 
usually even, such as 2.4.18) and the unstable one (the second number of the 
version is odd, such as 2.5.12). As ever, Torvalds is the person that decides 
what goes into which branch (many of the most controversial decisions are 
related precisely to this point). In any case, Linux does not have any planned 
deliveries in fixed timeframes: it will be ready when it is ready and in the 

GNUFDL» PID_001 48386 133 Free Software 

meantime we'll just have to wait. Surely by now, most readers will have realised 
that the decision on whether the system is ready or not will be made solely 
by Linus. 

The development method used in Linux has proven to be very effective in 
terms of results: Linux is very stable and any bugs are corrected extremely 
quickly (sometimes in minutes), as it has thousands of developers. In this sit- 
uation, when there is a bug, the probability that someone will find it is very 
high, and if the person that discovers it is not able to correct it, someone will 
appear who will hit on the solution very quickly. To summarise, this shows 
how Linux has thousands of people working on its development every month, 
which is why its success is not altogether surprising. 

It should be noted, however, that this way of working is very expensive where 
resources are concerned. It is not unusual for there to be many mutually-ex- 
clusive proposals for a new function or that a dozen patches are received for 
the same bug. In most cases, only one of the patches will finally be included 
in the kernel, which means that the rest of the time and effort put into the 
patches by the other developers will have all been in vain. Linux's develop- 
ment model is, therefore, a model that works very well in Linux but which 
not all projects can permit themselves. 

9.1.3. Linux's current status 

In early 2007, Linux was at version 2.6, which included, in terms of improve- 
ments made to version 2.4, NUMA (Non-Uniform Memory Access, used a 
lot in multiprocessors), new filesystems, improvements to communication in 
wireless networks and sound architectures (ALSA) and many other improve- 
ments (if you're interested in the details of the changes in respect of previous 
versions, you may consult "The wonderful world of Linux 2.6" [186]). 

Linux's development model has undergone some changes over recent years. 
Although the development mailing list is still the soul of the project, the code 
no longer has to pass through the list, necessarily. One of the things that have 
contributed to this in a large way is BitKeeper, a proprietary system that per- 
forms revision control, developed by the company BitMover, strictly following 
Linus Torvalds' recommendations. The use of this proprietary tool generated 
a lot of controversy, in which Linus, true to form, demonstrated his pragma- 
tism again, as for him and many others, the CVS version control system was 
obsolete. The disagreements were brought to an end with the development of 
git, a revision control system with similar features to BitKeeper that is current- 
ly used in Linux's development. More specifically, Linux's development pro- 
cess follows a pyramidal hierarchy, in which the developers propose patches, 
shared via mail between levels, which have to be accepted by the next level 

GNUFDL' PID 00148386 


Free Software 

up, formed by controller and file maintainers. The subsystem maintainers are 
on a higher level, whereas Linus Torvalds and Andrew Morton are on the top 
level and have the final say where the acceptance of patches is concerned. 

To summarise, the following table provides an x-ray picture of the Linux 
project, showing how it now has more than five million lines of code and that 
it can therefore be included amongst the largest free software projects (along 
with Mozilla and As to the estimates regarding the time it 
would take to design such a project and the average number of developers 
that would be necessary, we should note that the former is certainly much 
less than the time that Linux has been around. On the other hand, this is 
more than compensated by the latter detail, given that the average number 
of developers working full-time that would be necessary for such a project is 
higher than the number ever available to Linux. 


The cost estimate that COCOMO shows is in the range of 215 Million US Dollars, a sum 
that, if we put it in the context of everyday figures that we might think about, would be 
double what the best football clubs might pay for a great football star. 

Table 4. Analysis of Linux 


Beginning of the project 

First message on news.comp.os.minix: Au- 
gust 1991 



Analysed version 

2.6.20 (stable version on 20/02/2007) 

Source code lines. 


Cost estimate (according to basic COCO- 


Design time estimates (according to basic 

8.83 years (105.91 months) 

Estimate of average number of developers 
(according to basic COCOMO) 


Approximate number of developers 

These are estimated in the thousands (al- 
though only hundreds appear in the credits 

Development assistance tools 

Mailing list and git 

Linux's composition in terms of programming languages shows a clear pre- 
domination of C, which is considered to be an ideal language for designing 
speed-critical systems. When speed is such a strict requirement that not even 
C can achieve it, assembly language is directly used for programming and this, 
as we can see, happens with some frequency. The disadvantage of this assem- 
bly language, in comparison with C, is that it is not as portable. Each archi- 
tecture has its set of particular instructions, which means that a lot of code 

GNUFDL* PID 00148386 


Free Software 

written for an architecture in assembly language has to be ported to the oth- 
er architectures. The incidence of the rest of the languages, as shown in the 
attached table, is marginal and they are limited to installations functions and 
development utilities. The version analysed for this book was Linux 2.6.20, 

as it was published on 20* February 2007 (without including any subsequent 

Table 5. Programming languages used in Linux 

Programming language 

Code lines 

















9.2. FreeBSD 

As we have mentioned in the chapter on the history of free software, there are 
other types of free software operating systems, apart from the popular GNU/ 
Linux. A family of these are the "inheritors" of the distributions of Berkeley 
University, in California (US): BSD type systems. The oldest and most well- 
known BSD system is FreeBSD, which was created in early 1993, when Bill 
Jolitz stopped publishing the unofficial updates to 386BSD. With the assis- 
tance of the company Walnut Creek CDROM, which subsequently changed 
its name to BSDi, a group of volunteers decided to carry on creating this free 
operating system. 

The main objective of the FreeBSD project is the creation of an operating sys- 
tem that can be used without any type of obligations or ties, but that has all 
the advantages of code availability and is carefully processed to guarantee the 
quality of the product. The user has the liberty to do whatever they like with 
the software, either by modifying it according to their wishes or by redistribut- 
ing it in an open form or even in a closed form, under the terms that they 
wish, with or without modifications. As the name itself indicates, the FreeBSD 
project is based, therefore, on the philosophy of BSD licenses. 

9.2.1. History of FreeBSD 

Version 1.0 appeared towards the end of 1993 and was based on 4.3BSD Net/2 
and 386BSD. 4.3BSD Net/2 had code that was created in the seventies, when 
Unix was being developed by AT&T, which, as it turned out, involved a series 
of legal problems that were not resolved until 1995, when FreeBSD 2.0 was 
published without the original code developed by AT&T but based on 4.4BSD- 

GNUFDL" PID_001 48386 136 Free Software 

Lite, a light version of 4.4BSD (in which many of the modules had been elim- 
inated for legal reasons, apart from the fact that the port for Intel systems was 
still incomplete) that was released by the University of California. 

The history of FreeBSD would not be complete if we neglected to mention its 
"sister" distributions, NetBSD and OpenBSD. NetBSD appeared as version 0.8 
in the middle of 1993. The main aim was for it to be very portable (although at 
the beginning it was only an adaptation for i386); consequently, the product's 
motto was: "Of course it runs NetBSD". OpenBSD arose from the division of 
NetBSD caused by philosophical differences (as well as personal differences) 
between developers in mid-1996. The focus is mainly on security and cryptog- 
raphy and they say that it is the safest operating system that exists, although, 
as it is based on NetBSD, it is also highly portable. 

9.2.2. Development in FreeBSD 

The development model used by the FreeBSD project is based mainly on two 
tools: The CVS version control system and the GNATS bug-tracking software. 
The whole project is based on these two tools, as is confirmed by the fact that 
a hierarchy has been created on the basis of these tools. In effect, it is the 
committers (the developers with write-access to the CVS repository) who have 
the most authority for the project, either directly or indirectly through the 
choice of the core group, as we shall see in the next section. 

You do not have to be a committer in order to make bug reports in GNATS, 
which means that anyone who wishes to can report a bug. All the (open) con- 
tributions in GNATS are evaluated by a committer, who may assign the (anal- 
ysed) task to another committer or request more information from the person 
that originally made the bug report (feedback). There are situations in which 
the bug has been solved for some recent branches, which will then be speci- 
fied with the suspended status. In any case, the goal is that the report should 
be closed, once the error has been fixed. 

FreeBSD distributes its software in two forms: on the one hand, the ports, 
a system that downloads the source codes, compiles them and installs the 
application in the local computer, and on the other, the packages, which are 
simply the source codes of the precompiled ports and, therefore, in binary. The 
most important advantage of the ports over the packages is that the former 
allow the user to configure and optimise the software for their computer. On 
the other hand, in the package system, as they are already precompiled, it 
takes much less time to install the software. 

9.2.3. Decision-making process in FreeBSD 

The board of directors of FreeBSD, famously called the core team, is in charge of 
defining the direction of the project and ensuring that the objectives are met, 
as well as mediating in cases in which there are conflicts between committers. 

GNUFDL» PID_001 48386 137 Free Software 

Until October 2000, it was a closed group, which could only be joined by an 
express invitation from the core team itself. As of October 2000, the members 
are elected periodically and democratically by the committers. The most im- 
portant rule for the election of the core team is as follows: 

1) The committers that have made at least one commit over the last year have 
the right to vote. 

2) The Board of Directors will be renewed every two years. 

3) The members of the board of directors may be "expelled" by a vote of two 
thirds of the committers. 

4) If the number of members of the board of directors is less than seven, new 
elections will be held. 

5) New elections are held when a third of the committers vote for this. 

6) Any changes in the rules require a quorum of two thirds of the committers. 

9.2.4. Companies working around FreeBSD 

There are numerous companies that offer services and products based on 
FreeBSD, which FreeBSD lists on the project's website. In this presentation of 
FreeBSD we will learn more about the most significant aspects: BSDi and Wal- 
nut Creek CDROM. 

FreeBSD was born partly due to the foundation of the company BSDi in 1991 
by the people from CSRG (Computer Systems Research Group) of the Univer- 
sity of Berkeley, which would provide commercial support for the new oper- 
ating system. Apart from the commercial version of the FreeBSD operating 
system, BSDi also developed other products, such as an Internet server and a 
gateway server. 

Walnut Creek CDROM was created with the aim of commercialising FreeBSD 
as the final product, in such a way that it could be considered as a distribution 
in the style of those that exist for GNU/Linux, but with FreeBSD. In Novem- 
ber 1998, Walnut Creek broadened its horizons with the creation of the FreeB- 
SD Mall portal, which would commercialise all types of products based on 
FreeBSD (from the distribution itself to t-shirts, magazines, books, etc.), and 
announce third-party products on its website and provide professional FreeB- 
SD support. 

In March 2000, BSDi and Walnut Creek merged under the name BSDi to work 
together against the Linux phenomenon, which was clearly leaving BSD sys- 
tems in general and FreeBSD particularly, standing in the shadows. A year lat- 

GNUFDL» PID_001 48386 138 Free Software 

er, in May 2001, Wind River purchased the part that was dedicated to gener- 
ating the BSDi software, with the clear intention of boosting the development 
of FreeBSD for its use in embedded systems and intelligent devices connected 
to the Network. 

9.2.5. Current status of FreeBSD 

According to the latest data from the poll that Netcraft performs periodically, 
the number of web servers that run FreeBSD is approximately two million. 
A new user who wished to install FreeBSD could choose between version 6.2 
(which could be considered as the "stable" version) or the more advanced or 
"development" branch. Whilst the former provides more stability, especially 
in areas such as symmetric multiprocessing, which has been completely rede- 
veloped in the newer versions, the latter allows users to enjoy the latest de- 
velopments. It is also important to bear in mind that the developed versions 
tend to include test code, which slightly affects the system's speed. 

One of the star features of FreeBSD is what is known as the jails. The jails 
minimise the damage that might be caused by an attack on basic network 
services, such as Sendmail for the emails or BIND (Berkeley Internet Name 
Domain) for name management. The services are placed in a jail so that they 
run in an isolated environment. The jails can be managed using a series of 
utilities included in FreeBSD. 

9.2.6. X-ray picture of FreeBSD 

As we have mentioned in this last section, FreeBSD's functions are not restrict- 
ed solely to developing an operating system kernel, but also include the inte- 
gration of a multitude of utilities that are distributed together in the style of 
the GNU/Linux distributions. The fact that the development process of FreeB- 
SD is very closely linked to the CVS versions control system means that by 
studying the system, we can obtain a good idea of what FreeBSD consists of. 
The figures shown below correspond to the FreeBSD analysis performed on 

21 st August 2003. 

One of the most interesting aspects of FreeBSD is that the figures are very 
similar to the ones of in KDE and GNOME: the size of the software easily ex- 
ceeds five million lines of code, the number of files is approximately 250,000 
and the total number of commits is approximately two million. However, it 
is interesting to observe that the main difference between GNOME and KDE 
with respect to FreeBSD is the age of the project. FreeBSD recently made to its 
tenth year and it has been around for almost twice as long as the desktop en- 
vironments with which we are comparing it. That the size should be similar, 
despite the fact that the development period must have been longer, is partly 
due to the fact that FreeBSD did not attract as many developers. There is a list 
of approximately four hundred developers with write-access to the CVS (com- 
mitters), whereas the number of developers listed in the FreeBSD manual is 

GNUFDL' PID 00148386 


Free Software 

approximately one thousand. This is why the activity registered in FreeBSD's 
CVS is lower than the average (five hundred commits per day) than that reg- 
istered in both GNOME (nine hundred) and KDE (one thousand seven hun- 
dred, including the automatic commits). 

We have considered as the basic system of FreeBSD all that is placed in the 
src/src directory of the root module of the CVS. The activity that has been reg- 
istered in the basic system over the last ten years is that of more than half 
a million commits. There are more than five million lines of code, although 
we should remember that this does not only include the kernel, but many 
additional utilities, including games. If we take only the kernel into account 
(which is in the subdirectory sys), the scale is of 1.5 million of source code 
lines, predominantly in C. 

It is interesting to consider how the time estimate given by COCOMO corre- 
sponds exactly to the FreeBSD project's real time, although the estimate of the 
average number of developers is much higher than the effective number. We 
should also point out that in the last year, only seventy five committers have 
been active, whereas COCOMO supposes that over the ten years of develop- 
ment, the number of developers should be 235. 

Finally, we must remember, as we have mentioned, that FreeBSD's main ac- 
tivity is based around the CVS repository and the bug control system GNATS 

Table 6. Analysis of FreeBSD 


Beginning of the project 



Of BSD type 

Analysed version 


Source code lines. 


Lines of source code (kernel on 



Number of files 


Cost estimate 

$ 325,000,000 

Runtime estimate 

1 0.5 years (1 26 months) 

Estimate of average number of developers 


Approximate number of develo 


400 committers (1,000 collaborators) 

Number of committers active in 


last year 

75 (less than 20% of the total) 

Number of committers active in 


last two years 

1 65 (approximately 40% of the total) 

Number of commits in the CVS 


GNUFDL* PID 00148386 


Free Software 

Average number of commits (total) per day 

Approximately 500 

Development assistance tools 

CVS, GNATS, mailing list and news site 

C is the predominant language in FreeBSD and it keeps a greater distance from 
C++ than the other case that we have studied in this chapter. It is interesting 
to note that the number of lines of code in the assembly language contained 
in FreeBSD, matches, in terms of the scale, those of Linux, although those 
corresponding to the kernel are only twenty five thousand, in total. To sum- 
marise, we could say that in FreeBSD, the more classical languages within free 
software, C, Shell and Perl predominate, and that the other languages that we 
have looked at in other applications and projects, such as C++, Java, Python, 
have not been integrated. 

Table 7. Programming languages used in FreeBSD 

Programming language 

Code lines 












1 1 6,000 

1 .5% 



1 .20% 




9.2.7. Academic studies on FreeBDS 

Despite certainly being a very interesting project (we can look at its history by 
analysing the versions system, going back up to 10 years!), FreeBSD has not 
inspired that much interest in the scientific community. There is, however, 
one research team that has shown interest in the FreeBSD project, from vari- 
ous points of view ("Incremental and decentralised integration in FreeBSD") 
[149], which has especially focused on how software integration problems are 
resolved in an incremental and decentralised fashion. 

9.3. KDE 

Although it was not the first solution in terms of user-friendly desktop environ- 
ments, the dissemination of the Windows 95 operating system in mid-1995 
caused a radical change in the way non-specialised users interacted with com- 
puters. From the one-dimensional systems of lines of instructions (the termi- 
nals), the metaphor of the two-dimensional desktop environment was born, 
where the mouse began to be used more than the keyboard. Windows 95, 
more than a technological innovation, must be credited as being the system 

GNUFDL' PID 00148386 


Free Software 

that managed to cover all the personal and office environments, setting the 
standards that would be followed in the future (technical and social rules that, 

we are still, in some cases, suffering from in the early 21 st Century). 

Before desktop systems were created, each application managed its own ap- 
pearance and manner of interacting with the user, autonomously On desk- 
tops, however, the applications must have common properties and an appear- 
ance that is shared by the applications, which eases the pressure on the user, 
who can reuse the way of interacting learnt whilst using one application, when 
using another. This also eased the pressure on the application developers, as 
they did not have to deal with the problem of creating the interactive elements 
starting from zero (which is always a complicated task), but could start from 
a predefined framework and predefined rules. 

9.3.1. History of KDE 

Unix followers were quick to notice the outstanding success of Windows 95 
and, given that Unix-like environments did not have systems that were as 
intuitive whilst still being free, they decided to get to work. The KDE K Desktop 
Environment project was born from this effort in 1996; it was designed by 
Matthias Ettrich (creator of LyX, an editing program in the TeX typeset) and 
other hackers. The KDE Project proposed the following aims: 

• To provide Unix-like systems with a user-friendly environment that was, 
at the same time, open, stable, trustworthy and powerful. 

• To develop a set of libraries for writing standard applications on a graphical 
system for Unix XI 1. 

• To create a series of applications that would allow the user to achieve their 
objectives effectively and efficiently. 


Originally, the name KDE 
stood for Kool Desktop En- 
vironment, but it was subse- 
quently changed simply to K 
Desktop Environment. The of- 
ficial explanation was that the 
letter K is just before the L, for 
Linux, in the Roman alphabet. 

A controversy was created when the members of the newly-created KDE 
project decided to use an object-oriented library called Qt, belonging to the 
Norwegian firm Trolltech, which was not covered under any free software li- 
cense. It turned out that, although the KDE applications were licensed under 
GPL, they linked with this library, which meant that it was impossible to redis- 
tribute them. Consequently, one of the four freedoms established by Richard 
Stallman in the Free Software Manifesto was being violated [117]. As of ver- 
sion 2.0, Trolltech distributes Qt under a dual license that specifies that if the 
application that uses the library operates under the GPL, then the license valid 
for Qt is the GPL. Thanks to this, one of the most heated and hot-tempered 
debates in the world of free software had a happy ending. 

GNUFDL" PID_001 48386 142 Free Software 

9.3.2. Development of KDE 

KDE is one of the few free software projects that generally follows a new ver- 
sion launch schedule (let us remember, for example, that there will be a new 
Linux version "when it is ready", whereas, as we shall discuss later, GNOME 
has always suffered significant delays when it came to releasing new versions). 
The numbering of the new versions follows a perfectly defined policy. The 
KDE versions have three version numbers: a higher one and two lower ones. 
For example, in KDE 3.1.2, the higher number is the 3, whereas the 1 and 2 
are the lower numbers. Versions with the same higher number have binary 
compatibility, which means that it is not necessary to recompile the applica- 
tions. Until now, the changes in the higher number occurred in parallel with 
the changes in the Qt library, which shows how the developers wanted to 
take advantage of the new functionalities in the Qt library in the imminent 
version of KDE. 

Where the lower numbers are concerned, the versions with one single lower 
number are versions in which they have included both the new functionali- 
ties and in which the bugs that have been discovered, have been corrected. 
The versions with a second lower number do not include new functionalities 
in respect of the versions with the first lower number, and only contain the 
bug corrections. The following example will explain this better: KDE 3.1 is a 
third-generation version of KDE (higher number 3) to which new functionali- 
ties have been added, whereas KDE 3.1.1 is the previous version with the same 
functionalities, but with all the bugs that have been found corrected. 

KDE was built, shortly after the project began, in an association registered in 
Germany (KDE e.V.) and, as such, the articles of association meant that there 
has to be a managing committee. The influence of this managing committee 
on the development is nil, as its main task is the administration of the associ- 
ation, especially where the donations that the project receives are concerned. 
In order to promote and disseminate KDE, the KDE League, which includes 
all interested companies, was created, as we shall discuss below. 

9.3.3. The KDE League 

The KDE League is a group of companies and individuals from KDE that have 
the objective of enabling the promotion, distribution and development of 
KDE. The companies and individuals that participate in the KDE League do 
not have to be directly involved in the development of KDE (although the 
members are encouraged to get involved), but simply represent an industrial 
and social framework that is friendly to KDE. The aims of the KDE League are 
as follows: 

• Promoting, providing and facilitating the formal and informal education 
of the functionalities, capabilities and other qualities of KDE. 

GNUFDL» PID_001 48386 143 Free Software 

• To encourage corporations, governments, companies and individuals to 
use KDE. 

• To encourage corporations, governments, companies and individuals to 
participate in the development of KDE. 

• To provide knowledge, information, management and positioning around 
KDE in terms of its use and development. 

• To foster communication and cooperation between KDE developers. 

• To foster communication and cooperation between KDE developers and 
the general public through publications, articles, websites, meetings, par- 
ticipation in conferences and exhibitions, press articles, interviews, pro- 
motional materials and committees. 

The companies that participate in the KDE League are mainly distribution de- 
signers (SuSE, now part of Novell, Mandriva, TurboLinux, Lindows and Han- 
com, a Korean free software distribution), development companies (Trolltech 
and Klaralvdalens Datakonsult AB), plus the giant IBM and a company created 
with the aim of promoting KDE ( Of all these, we must especially 
mention Trolltech, Novell and Mandriva Software, whose involvement has 
been essential and whose business models are very closely linked to the KDE 

• Trolltech is a Norwegian company based in Oslo that develops Qt, the li- 
brary that can be used as a graphic interface for the user and an API for 
developers, although it can also work as an element embedded in PDA 
(such as the Sharp Zaurus). The importance of the KDE project for Troll- 
tech is evidenced by two basic elements in its commercial strategy: on 
the one hand, it recognises KDE as its main promotion method, encour- 
aging the development of the desktop and accepting and implementing 
the proposed improvements or modifications; on the other hand, some of 
the most important KDE developers work professionally for Trolltech; the 
most well-known example is that of Matthias Ettrich himself, who found- 
ed the project, which doubtlessly benefits both the KDE project and the 
company itself. Trolltech's involvement in the KDE project is not exclu- 
sively limited to the Qt library, as is proven by the fact that one of the 
main developers of KOffice, KDE's office software package, currently has 
a part-time contract with them. 

• SuSE (now part of Novell) has always shown its special predilection for 
the KDE desktop system, partly due to the fact that most of its developers 
are of German or Central European origin, as is the company itself. SuSE, 
aware of the fact that the better and easier the desktop environment that 
its distribution offers, the greater its implementation and, therefore, the 
sales and support requests, has always had a very active policy in terms 

GNUFDL» PID_001 48386 144 Free Software 

of the budget allocated to professionalising key positions within the KDE 
project. As an example, the current administrator of the version control 
system and another two of the main developers are all on SuSE's payroll. 
Likewise, within SuSE's workforce there are a dozen developers that can 
spend some of their working time on developing KDE. 

• The Mandriva distribution is another one of the biggest backers of KDE 
and a number of the main developers work for it. Its financial situation, 
which has included bankruptcy from 2003, has meant that it has lost in- 
fluence over the last few years. 

9.3.4. Current status of KDE 

After the publication of KDE 3 in May 2002, the general opinion is that the free 
desktops are on a par with their proprietary competitors. Some of its greatest 
achievements include the incorporation of a components system (KParts) that 
makes it possible to embed some applications in others (a piece of a KSpread 
spreadsheet in the KWord word processor) and the development of DCOP, a 
simple system for processes to communicate with each other, with authenti- 
cation. DCOP was the project's commitment that acted in detriment to the 
CORBA technologies, a widely-debated subject within the world of free desk- 
tops, especially for GNOME, where it was decided that CORBA would be used. 
History seems to have put each technology in its place, as can be seen from the 
DBUS proposal (an improved type of DCOP) from, a project 
interested in promoting the interoperability and the use of joined technolo- 
gies in free desktops, which is, coincidentally, led by one of the most well- 
known GNOME hackers. 

The following table summarises the most important characteristics of the KDE 
project. The licenses that the project accepts depend on whether they are for 
an application or a library. The library licenses provide greater "flexibility" for 
third parties; in other words, they make it possible for third parties to create 
proprietary applications that are linked to the libraries. 

The latest KDE version is, as at early 2007, version 3.5.6 and the fourth gener- 
ation, KDE 4, which will be based on Qt4, is expected to arrive in mid-2007. 
The generation change involves a lot of effort on adapting the version, which 
is a tedious and time-consuming task. However, this does not mean that the 
"old" applications will no longer work. Generally, in order to having them still 
working, the older versions of the libraries on which they were based are also 
included, although this means that various versions of the libraries have to 
be loaded simultaneously in the memory, with the ensuing waste of system 
resources. The KDE developers view this effect as an inherent part of the de- 
velopment of the project and, therefore, as a lesser evil. 

GNUFDL' PID 00148386 


Free Software 

9.3.5. X-ray picture of KDE 

Where the scale of KDE is concerned, the figures that we will now discuss 
correspond to the status of CVS in August 2003, which means that they should 
be taken with the usual reservations that we have already discussed, plus one 
more: some of the modules that have been used in this study are still under 
development and do not fulfil the criteria of being a finished product. This 
shouldn't really have any effect for our purposes, as we are more interested in 
the scale of the results than the exact numbers. 

The source code included in KDE's CVS is in the total sum of six million lines 
of code in different programming languages, as we shall show below. The time 
required to create KDE would be approximately nine and a half years, which 
is more than the project's seven years, and the estimated average number of 
developers working full-time would be two hundred. If we take into account 
the fact that KDE had approximately eight hundred people with write-access 
to CVS in 2003 (of which half have been inactive over the last two years) 
and the fact that the number of KDE developers with full-time contracts has 
not been more than twenty at any given time, we can see that KDE's level of 
productivity is much, much higher than the estimate provided by COCOMO. 


A company that wanted to develop a product of this scale starting from zero would have 
to invest more than 250 million dollars; for comparative purposes, this sum would be 
equivalent to an investment of a car manufacturer in the creation of a new production 
plant in Eastern Europe or what a well-known oil company is planning to spend in order 
to open two hundred petrol stations in Spain. 

It is also interesting to see that a large part of the effort, almost half of that 
expended on the development of the KDE project, would correspond to the 
translation of the user interface and the documentation. Although very few 
(approx. one thousand) of the programming lines are concerned with this 
task, the number of files dedicated to this purpose is seventy five thousand 
for translations (a sum that increases to one hundred thousand if we include 
the documentation in the different formats), which comprises almost a fourth 
(third) of the 310,000 files that there are in CVS. The combined activity of 
CVS is of one thousand two hundred commits per day, which means that the 
average time between commits is approximately one minute 10 . 

( Two observations should be 
made on this point: the first is that 
when a commit including various 
files is made, it is as if there had 
been one separate commit for each 
file; the second is that the number 
of commits is an estimated sum, as 
the project has a series of scripts 
that perform commits automatical- 

Where the tools, the information locations and the development assistance 
events are concerned, we will see that the range of possibilities offered by KDE 
is much wider than that used in Linux. Apart from the version control system 
and the mailing lists, KDE has a series of websites providing information and 
technical and non-technical documentation on the project. There is also a 
news site among these sites where new solutions are presented and proposals 
are debated. The news site, however, cannot be considered as a replacement 
for the mailing lists, which, as occurs with Linux, is where the real debates 
take place and the decisions are made and the future strategies devised; the 

GNUFDL' PID 00148386 


Free Software 

news site is really more of a meeting point for the users. Finally, KDE has been 
organising regular meetings for three years, in which the developers and the 
collaborators meet for approximately a week to present the latest innovations, 
develop, debate and get to know each other and have a good time (not nec- 
essarily in that order). 

Table 8. KDE Analysis 


Beginning of the project 


License (for applications) 

GPL, QPL, MIT, Artistic 

License (for libraries) 

LGPL, BSD, X1 1 

Analysed version 


Source code lines. 


Number of files (code, documentation, etc.) 

31 0,000 files 

Cost estimate 

$ 255,000,000 

Runtime estimate 

9.41 years (112.98 months) 

Estimate of average number of developers 


Approximate number of developers 

Approximately 900 committers 

Number of committers active in the last year 

Around 450 (approximately 50% of the to- 

Number of committers active in the last two 

Around 600 (approximately 65% of the to- 

Approximate number of translators (active) 

Approximately 300 translators for more than 
50 languages (including Esperanto). 

Number of commits (by developers) in the 

Approximately 2,000,000 (estimated figure 
not including automatic commits) 

Number of commits (by translators) in the 

Approximately 1,000,000 (estimated figure 
not including automatic commits) 

Average number of commits (total) per day 


Tools, documentation and development as- 
sistance events 

CVS, mailing lists, website, news site, annual 

Where the programming languages used in KDE are concerned, C++ predom- 
inates. This is mainly due to the fact that this is the native language of Qt, 
although a great effort is expended on providing bindings to allow develop- 
ments in other programming languages. Certainly, the number of lines of code 
in the minority languages corresponds almost integrally to those bindings, al- 
though this does not mean that they are not used at all, as there are numerous 
projects external to KDE that use them. 

GNUFDL* PID 00148386 


Free Software 

Table 9. Programming languages used in KDE 



Code lines 








Objective C 





1 .69% 



1 .44% 




9.4. GNOME 

The main objective of the GNOME project is to create a desktop system for the 
end user that is complete, free and easy to use. Likewise, the idea is for GNOME 
to be a very powerful platform for developers. The initials GNOME stand for 
GNU Network Object Model Environment. From its name, we see that GNOME is 
part of the GNU project. Currently all the code contained in GNOME must be 
under a GNU GPL or a GNU LGPL license. We can also see that the networks 
and the object-orientated modelling are extremely important. 

9.4.1. History of GNOME 

Whilst the freedom of KDE was still being argued about, in the summer of 
1997, as fate would have it, Miguel de Icaza and Nat Friedman met at Redmond 
during some workshops organised by Microsoft. It is probable that this meet- 
ing caused a radical turnaround in both, resulting in the creation of GNOME 
by Miguel de Icaza when he returned to Mexico (along with Federico Mena 
Quintero) and his admiration for distributed object technologies. De Icaza and 
Mena decided to create an environment that would be an alternative to KDE, 
as they understood that a reimplementation of a proprietary library would 
have been a task destined to failure. And thus GNOME was born. 

Since those ancient times in 1997, GNOME has gradually grown and con- 
tinues to grow, with its repeated publications. Version 0.99 was launched in 
November 1998, but the first really popular version, distributed practically 
with any GNU/Linux distribution, would be GNOME 1.0, published in March 
1999. It should be noted that the experience of this first stable version of 
GNOME was not very satisfactory as many considered it to be full of critical 
bugs. For this reason, GNOME October (GNOME 1.0.55) is treated as the first 
version of the GNOME desktop environment that was truly stable. As we can 
observe, with GNOME October, the developers did not use numerated publi- 
cation version so as to avoid entering a "versions race" against KDE. The first 
GUADEC, the GNOME users and developers European conference, was held 
in Paris in 2000 and narrowly missed coinciding with the publication of the 

GNUFDL» PID_001 48386 148 Free Software 

new version of GNOME, called GNOME April. It was the last version to be 
named after a month, as it turned out that this system created more confusion 
than anything else (for example, GNOME April was launched after GNOME 
October, although one could be forgiven for assuming the opposite). In Octo- 
ber of that year, after numerous debates over the months in different mailing 
lists, the GNOME Foundation, which we shall discuss in subsequent sections, 
was created. 

GNOME 1.2 was a step forward in terms of the architecture used by GNOME, 
an architecture that continued to be used in GNOME 1.4. This era was charac- 
terised by the second GUADEC, which took place in Copenhagen. What had 
begun as a small meeting of a few hackers, became a great event that captured 
the attention of the whole software industry. 

In the meantime, the argument about the freedom of KDE was resolved with 
Trolltech's change of position, when it ended up licensing Qt under a dual 
license, which was for free software for applications that work with free soft- 
ware. Today, there is no doubt that both GNOME and KDE are free desktop 
environments, which means that we can say that the development of GNOME 
has encouraged the creation of not just one free desktop environment, but 

9.4.2. The GNOME Foundation 

The most difficult problem to take on board when you hear about GNOME for 
the first time is the organisation of the more than one thousand contributors 
to the project. It is paradoxical that a project with a structure that tends toward 
the anarchic, should be this successful and achieve complex objectives that 
only a few multinationals in the IT sector would be able to achieve. 

Although GNOME was created with the clear aim of providing a user-friend- 
ly and powerful environment, to which new programs would gradually be 
added, it soon became apparent that it would be necessary to create a body 
that would have certain responsibilities that would allow them to promote 
and boost the use, development and dissemination of GNOME: consequently, 
the GNOME Foundation was created in 2000; its headquarters are situated in 
Boston, US. 

The GNOME Foundation is a non-profit organisation and not an industrial 
consortium; it has the following functions: 

• Coordinating the publications. 

• Deciding which projects are part of GNOME. 

GNUFDL» PID_001 48386 149 Free Software 

• It is the official spokesperson (for the press and for both commercial and 
non-commercial organisations) of the GNOME project. 

• Promoting conferences related to GNOME (such as the GUADEC). 

• Representing GNOME in other conferences. 

• Creating technical standards. 

• Promoting the use and development of GNOME. 

In addition, the GNOME Foundation receives financial funds for promoting 
and boosting the functions mentioned above, as this was impossible to do in 
a transparent manner before the foundation was created. 

Currently, the GNOME Foundation has one full-time employee that is in 
charge of solving all the bureaucratic and organisational tasks that have to 
be done in a non-profit organisation that holds regular meetings and confer- 

In general terms, the GNOME Foundation is divided into two large commit- 
tees: a managing committee and an advising committee. 

The managing committee (the Board of Directors) is formed, at the most, by 
fourteen members elected democratically by the members of the GNOME 
Foundation. A "meritocratic" model is followed, which means that, in order 
to be a member of the GNOME Foundation, one has to have cooperated in 
one way or another with the GNOME project. The contribution does not nec- 
essarily have to involve source code; there are also tasks that require transla- 
tion, organisation, dissemination, etc., which one could perform and then ap- 
ply for membership of the GNOME Foundation, in order to have the right to 
vote. Therefore, it is the members of the Foundation that can put themselves 
forward for the board of directors and it is the members that, democratical- 
ly, choose their representatives on the board from the persons that have put 
themselves forward. Currently, voting is by email. The duration of the term 
as member of the board of directors is one year, after which elections are held 

There are some basic regulations for guaranteeing the transparency of the 
board of directors. The most remarkable one is the limitation on the number 
of members affiliated to the same company, which cannot exceed four em- 
ployees. It is important to emphasise that the members of the board of direc- 
tors are always so in their personal capacity, and never in representation of a 
company. Nevertheless, after a long discussion, it was agreed that this clause 
would be included to avoid any mistrust. 

GNUFDL» PID_001 48386 ISO Free Software 

The other committee within the GNOME Foundation is the advising commit- 
tee, which has no authority to make decisions but that serves as a vehicle for 
communicating with the managing committee. It is formed by commercial 
companies working in the software industry, as well as non-commercial or- 
ganisations. Currently, its members include Red Hat, Novell, Hewlett-Packard, 
Mandrake, SUN Microsystems, Red Flag Linux, Wipro, Debian and the Free 
Software Foundation. All companies with more than ten employees are re- 
quired to pay a fee in order to be part of the board of advisors. 

9.4.3. The industry working around GNOME 

GNOME has managed to enter the industry substantially, in such a way that 
various companies have participated very actively in its development. Of all of 
these, the most important cases are those of Ximian Inc., Eazel, the RHAD Labs 
by Red Hat and, more recently, SUN Microsystems. We will now describe, for 
each case, the motivations of the companies as well as their most important 
contributions to the GNOME desktop environment: 

• Ximian Inc. (originally called Helix Inc.) is the name of the company that 
was founded in 1999 by Miguel de Icaza, the cofounder of GNOME, and 
Nat Friedman, one of GNOME's hackers. The main objective was to bring 
together the most important GNOME developers under the same umbrel- 
la to maximise development, which is why it is not surprising that its 
current and past employees have included around twenty of the most ac- 
tive GNOME developers. The application that Ximian put the most effort 
into from the very start was Evolution, a complete personal information 
management system in the style of Microsoft Outlook, which included an 
email client, agenda and a contacts address book. The products that Ximi- 
an sold were the Ximian Desktop (a version of GNOME with more corpo- 
rate purposes), Red Carpet (mainly, although not limited to, a GNOME 
software distribution system) and finally MONO (a reimplementation of 
the .NET development platform), although the latter project is not, as yet, 
related in any way to GNOME. Ximian also developed an application that 
permits Evolution to interact with an Exchange 2000 server. This applica- 
tion, whilst being quite small, was very controversial because it was pub- 
lished with a non-free license (subsequently, in 2004, this component was 
also licensed as free software). In August 2003, Novell, as part of its strat- 
egy for entering the GNU/Linux desktop, bought Ximian. 

• Eazel was founded in 1999 by a group of people who used to work for 
Apple, with the aim of making the GNU/Linux environment as easy as 
the Macintosh environment. The application on which they concentrated 
their efforts was called Nautilus and it was supposed to be the file manager 
that would definitively retire the mythical Midnight Commander, devel- 
oped by Miguel de Icaza. The lack of a business model and the dotcoms 
crisis, which caused risk investors to remove all the capital that was re- 
quired for the company to carry on working, resulted in Eazel declaring 

GNUFDL» PID_001 48386 151 Free Software 

bankruptcy on 15 4 May 2001 and closing its doors. It did have time to 
release Nautilus version 1.0 before this however, although the numbering 
was rather artificial, given that the stability that one would expect in a 
1.0 version was nowhere to be seen. Two years after Eazel's bankruptcy, 
we were able to see how Nautilus had developed and become a complete 
and manageable file manager integrated in GNOME; this means that the 
story of Eazel and Nautilus can be considered as a paradigmatic case of a 
program that survives the disappearance of the company that created it; 
something that is almost only possible in the world of free software. 

• Red Hat created the Red Hat Advanced Development Labs, RHAD, with 
the aim of ensuring that the GNOME desktop would gain user-friendliness 
and power. In order to achieve this, Red Hat used half a dozen of the most 
important hackers from GNOME and gave them the freedom to develop 
whatever they decided was appropriate. From the RHAD Labs we have 
ORBit, the implementation of CORBA used by the GNOME project, known 
as "the fastest in the west". Another important aspect is the work that was 
carried out on the new version of GTK+ and on GNOME's configuration 
system, GConf. 

• SUN Microsystems became involved in the development of GNOME at a 
later stage, as GNOME was a relatively mature product by September 2000. 
SUN's intention was to use GNOME as the desktop system for the Solaris 
operating system. It therefore created a team to work with GNOME, whose 
most important merits include the usability and accessibility of GNOME. 
In June 2003, SUN announced that it would distribute GNOME 2.2 with 
version 9 of Solaris. 

9.4.4. GNOME's current status 

GNOME, as at early 2007, is at version 2.18. Most of the key technologies 
on which it is based have matured, as is evident from the version number. 
For example, the CORBA broker used now is ORBit2, whilst the graphical en- 
vironment and API, GTK+, underwent changes devised from the experience 
accumulated during the previous versions of GNOME. One important novel- 
ty is the inclusion of an accessibility library, proposed by SUN, which allows 
people with accessibility problems to use the GNOME environment. A special 
mention should also go to Bonobo, the GNOME components system. Bonobo 
left its mark on an era within GNOME, whilst the personal information man- 
agement program Evolution was being developed. However, time has proven 
that the expectations raised by Bonobo were too high and that the reuse of the 
effort expended on it by employing its components has not been as extensive 
as was initially expected. 

GNUFDL» PID_001 48386 152 Free Software 


The ATK library is a library of abstract classes that makes applications accessible. This 
means that people with some form of disability (the blind, the colour-blind, people with 
eye problems, those who cannot use a mouse, a keyboard, etc.) may still use GNOME. 
SUN's interest on ensuring accessibility is due to the fact that if it wishes to offer its 
services to the government of the United States, it has to meet a series of accessibility 
standards. They have taken this work so seriously that there is even a blind programmer 
in the GNOME development team working at SUN. In September 2002, GNOME's acces- 
sibility architecture was given the Helen Keller Achievement Award. 

9.4.5. X-ray picture of GNOME 

The data and figures shown in table 10 bring us to the end of our presentation 

of GNOME. The figures correspond to the status of GNOME's CVS as at 14 th 
August 2003. On that date, there were more than nine million lines of code 
hosted in the CVS repository owned by the GNOME project. Even though 
the most natural thing would be to compare GNOME to KDE, we must warn 
readers that the differences in terms of how these projects are organised make 
this unadvisable if we wish to make the comparison in equal conditions. This 
is due, for example, to the fact that GNOME's CVS includes GIMP (a program 
for creating and handling graphics), which, on its own, represents more than 
660,000 lines of code, or the GTK+ library, on which GNOME's development 
focuses, and which, on its own, has 330,000 lines. If we add to this the fact that 
GNOME's CVS repository is more inclined to open new modules for programs 
(it has a total of seven hundred) than KDE's (which has less than one hundred), 
we can understand why GNOME has more lines than KDE, despite being a 
year and a half younger. The GNOME repository hosts more than 225,000 
files, which have been added and modified almost two million times (see the 
number of commits some rows below, in the table). 


A company that wanted to create software of the size of GNOME's software, would have 
to contract an average of approximately two hundred and fifty developers for more than 
eleven years, in order to obtain a product with a similar extension, according to the CO- 
COMO model used throughout this chapter. The associated cost would be approximately 
400 million dollars, a figure similar to that which a well-established mobile telephone 
company invested in 2003 to reinforce its network capacity, or similar to the figure that 
a car manufacturing firm would pay in order to open a production plant in Barcelona. 

GNOME's human resources include almost one thousand developers with write-access to 
the CVS revision control system, of which almost twenty work for GNOME professionally 
(full-time or part-time). Of these, only 25% have been active in the last year and 40% have 
been active over the last two years. The average number of commits per day, registered 
since the project began is almost one thousand. The development assistance tools used 
by the GNOME project are basically the same as those used by KDE, and so we will not 
go into them in this section. 

Table 10. Analysis of GNOME 


Beginning of the project 

September 1997 



GNUFDL* PID 00148386 


Free Software 

Analysed version 


Source code lines. 


Number of files (code, documentation, 


Cost estimate 

$ 400,000,000 

Runtime estimate 

1 1 .08 years (1 33.02 months) 

Estimate of average number of developers 

Approximately 250 

Number of subprojects 

More than 700 modules in the CVS. 

Approximate number of developers 

Almost 1 ,000 with write-access to the CVS. 

Number of committers active in the last 

Around 500 (approximately 55% of the total) 

Number of committers active in the last two 

Approximately 700 (75% of the total) 

Number of commits in the CVS 


Average number of commits (total) per day 

Approximately 900 

Development assistance tools 

CVS, mailing lists, website, news site, annual 

Whereas in KDE, C++ is undoubtedly the most widely-used language, in 
GNOME, the language is C. In GNOME, as occurs in KDE, this is due to the fact 
that the main library is written in C, which means that the native language is 
C, whereas programmers wishing to use the other languages have to wait for 
the corresponding bindings to appear. The most advanced language binding 
in GNOME is the one that is included in gnome--, which is none other than 
C++, which is why it is not surprising that that is the second language in the 
classification. Perl has always been widely accepted within the GNOME com- 
munity and an example of this fact is that in GNOME it is possible to program 
in many languages. Its implementation, however, has not been as extensive 
as could have been expected and it is slightly more extensive than Shell. On 
another note, Python and Lisp were accepted fairly widely in GNOME, as is 
proven by the relative importance of this classification, whereas Java has nev- 
er really taken off probably due to an incomplete link. 

Table 11. Programming languages used in GNOME 

Programming language 

Code lines 














GNUFDL* PID 00148386 154 Free Software 

Programming language 

Code lines 




1 .49% 




9.4.6. Academic studies on GNOME 

The most important studies on GNOME in the academic sphere are the fol- 
lowing two: "Results from software engineering research into open source de- 
velopment projects using public data" [158] and "The evolution of GNOME" 

• [158] is one of the first large-scale studies of software in the sphere of 
free software. The authors of the study took advantage of the fact that 
the details of the development are usually publicly accessible in order to 
measure the efforts and compare them against the cost estimate models, 
and traditional time and effort measurements. One of the classical models 
with which they compared them was the one used in this chapter, model 

• [123] briefly goes over the objectives of GNOME and its short history, as 
well as the GNOME project's use of technology. 

9.5. Apache 

The HTTP Apache server is one of the star applications of the world of 
free software, as it is the web server that is most widely used, according to 
the Netcraft real-time survey ( 
august_2003_web_server_survey.html) [167]. For example, in May 1999, 57% 
of web servers worked with Apache, whereas in May 2003, the percentage had 
increased to 68%. Apache is available for all types of Unix (BSD, GNU/Linux, 
Solaris...), Microsoft Windows and other minority platforms. 

9.5.1. History of Apache 

In March 1989, Tim Berners Lee, an English scientist that worked in the CERN 
(Switzerland) proposed a new method for managing the huge amount of in- 
formation from the CERN projects. The method would be a network of hy- 
perlinked documents (hypertext, as Ted Nelson had called it already in 1965); 
the WWW was born. However, it was not until November 1990 that the first 
WWW software was unveiled: a package called the World Wide Web included 
a web browser with a graphic interface and a WYSIWYG ("what you see is what 
you get") editor. Two years later, the list of WWW servers had approximately 
thirty entries, including NCSA HTTPd. 

GNUFDL» PID_001 48386 155 Free Software 

The real history of Apache began when Rob Mc Cool left the NCSA in March 

1995. Apache 0.2 would be born on 18 th March 1995, based on the NCSA 
HTTPd 1.3 server, built by Rob McCool himself while he was at NCSA. During 
those first months, Apache was a collection of patches applied to the NCSA 
server, until Robert Thau launches Shambhala 0.1, an almost complete reim- 
plementation that already included the API for the modules that subsequently 
turned out to be so successful. 


The name of the Apache project is based on its philosophy of development and organ- 
isation. As was the case with the Apache tribe, the developers of Apache decided that 
their organisational method should be based on the personal merits of the developers 
in comparison with the rest of the Apache community. However, there is a legend that 
has spread that says that the name Apache really came from the fact that in the initial 
stages, it was simply a patched NCSA server, or a patchy server. 

The first stable version of Apache did not appear until January 1996, when 
Apache 1.0 was released, which included the loading of modules in test-mode 
runtime, as well as other interesting functions. The first months of that year 
were especially fruitful for the project, as version 1.1, which had authentica- 
tion modules that would be checked against the databases (such as MySQL) 
was published only two moths later. From that time to today, the most im- 
portant events for the project have been the introduction of total compliance 
with the HTTP 1.1 standard (included in April 1997 in Apache 1.2), the inclu- 
sion of the Windows NT platform (which began in July 1997 with the test 
versions of Apache 1.3), the unification of the configuration files in one sin- 
gle file (which did not happen until October 1998, in Apache 1.3.3) and the 
launch, still in the test stage, of the next generation of Apache, Apache 2. 

In the meantime, in June 1998, IBM decided that, instead of developing its 
own HTTP, it would use Apache as the engine of its product WebSphere. This 
was interpreted as a huge endorsement for the Apache project from the Big 
Blue and for free software in general, although it would be necessary to modify 
the original Apache license slightly in order to make this work. 

9.5.2. Development of Apache 

The HTTP Apache server is the main project among the many that the Apache 
Software Foundation manages. The modular design of Apache has made it 
possible for there to be a series of satellite projects, based around Apache, some 
of which have even been bigger than Apache itself. For example, the HTTP 
Apache server contains the kernel of the system with the basic functionalities, 
whereas the additional functionalities are provided by different modules. The 
most well-known modules are mod_perl (a Perl script language interpreter em- 
bedded in the web server) and Jakarta (a powerful applications server). In the 

GNUFDL" PID_001 48386 156 Free Software 

following paragraphs, we will only describe the development process followed 
for the HTTP server, without taking into account the other modules, which 
may have similar processes or not. 

The development of the HTTP Apache server is based on the work of a small 
group of developers called the Apache Group. The Apache Group consists of 
the developers that have worked together on the project for a long period of 
time, generally more than six months. The developer, having been invited by 
a member of the Apache Group to join, is voted in by all the other members. In 
the early stages, the Apache Group consisted of eight developers; this number 
then increased to twelve and there are currently twenty five members. 

The Apache Group is responsible for the development of the web server and, 
therefore, for specific decisions regarding the development at any given mo- 
ment. It is important to distinguish the Apache Group from the developers in 
the core group, which is active at all times. The voluntary nature of the work 
performed by most of the developers makes it unlikely that all the people that 
comprise the Apache Group will be active at all times, which means that the 
core is defined as the people who may take care of the tasks in Apache in a 
certain period of time. In general, the decisions that have to be made by the 
developers belonging to the core group are limited to voting for the inclusion 
or not of code, although in reality this is reserved only for large-scale changes 
and questions of design. On another note, they usually have write-access to 
the CVS repository, which means that they act as guardians for the incoming 
code, ensuring that it is correct and of good quality. 

9.5.3. X-ray picture of Apache 

The figures shown below correspond to the HTTP Apache server version that 
was available for download from the CVS server of the Apache project as at 

18* April 2003. None of the numerous modules that the Apache project has, 
have been taken into account here. As we will see, the Apache project is rela- 
tively small as compared with the other cases studied in this chapter. Although 
this has already been mentioned, it is important to emphasise the modularity 
of Apache, which has the following specific advantages: the kernel is small 
and manageable. The CVS repository of the Apache project, which contains 
the kernel of the web server and many additional modules, hosts more than 
four million lines of source code, a figure that is slightly lower than those of 
projects such as KDE and GNOME. 

Version 1.3 of Apache had a little more than 85,000 lines of source code; ac- 
cording to the COCOMO model, this would have required the work of an av- 
erage of twenty developers working full-time for a year and a half. The total 
cost of the project would, at the time, be approximately 4 million dollars. In 

GNUFDL* PID 00148386 


Free Software 

order to prepare the Apache web server, up to sixty different committers would 
have been necessary, whereas the number of developers providing input, ac- 
cording to the calculations, would have been approximately four hundred. 

Table 12. Analysis of Apache 


Beginning of the project 



Apache Free Software License 

Analysed version 


Source code lines. 


Number of files 


Cost estimate 

$ 7,971,958 

Runtime estimate 

2.52 years (30.27 months) 

Estimate of average number of developers 


Approximate number of developers 

60 commiters (400 developers) 

Development assistance tools 

CVS, mailing lists, bug report system 

Apache 1.3 is written almost completely in C language and there are scarcely 
any other programming languages, especially if we take into account the fact 
that most of the lines written in the second language, Shell, correspond to 
configuration files and compilation assistance. 

Table 13. Programming languages used in Apache 

Programming language 

Code lines 














9.6. Mozilla 

The Mozilla project works on a set of integrated applications for Internet, that 
are free and multiplatform, and the most notable products are the Mozilla 
Firefox web browser and the Mozilla Thunderbird email and news client. This 
set is also designed as a platform for developing other applications, which 
means that there are many browsers that use Gecko, Mozilla's HTML engine 
(such as Galeon). 

GNUFDL» PID_001 48386 158 Free Software 

The project is managed by the Mozilla Foundation, a non-profit organisation 
that creates free software and is "dedicated to preserving choice and promoting 
innovation on the Internet". For this reason, Mozilla's products are based on 
three basic principles: they must be free software, respect the standards and 
be portable to other platforms. 

9.6.1. History of Mozilla 

The history of Mozilla is long and convoluted but also very interesting, as it 
allows us to follow the history of the WWW itself. The reason for this is that 
if we trace all the persons and institutions that have been involved in the 
development of Mozilla, then we arrive at the starting point of the Internet, 
with the launch of the first complete Internet browser. 

As was the case with Apache's predecessor, it was the NCSA where the first 
complete Internet browser, Mosaic, was "born" in 1993. Many of the members 
of the development team, with Marc Andreessen and Jim Clark at the helm, 
created a small company in order to write, starting from zero (as there were 
problems with the copyright on the code of Mosaic and the technical design of 
the program had its limitations, see Speeding the Net: the inside story of Netscape 
and how it challenged Microsoft [189]), what would subsequently become the 
Netscape Communicator browser, which was, unarguably, the leader of the 
market of Internet browsers until the arrival of Microsoft Internet Explorer. 
Apart from the purely technological innovation that the Netscape browser 
represented, Netscape Inc. was also innovative in the way it managed to cor- 
ner the market. Completely contrary to what was held as common sense at 
the time, its star application, the WWW browser, was available for free (and 
could even be distributed with certain limitations). This approach, which was 
completely unheard of in the corporate world at the time, caused a certain 
amount of surprise, but it turned out to be right for Netscape Inc.'s strategy, 
and it was only the giant Microsoft that was able to outdo it with more ag- 
gressive (and probably detrimental to free market competition) tactics. 

Around 1997, Netscape's market share had dropped sharply due to the spread 
of Microsoft Explorer; consequently, Netscape Inc. was studying new ways of 
recovering its previous dominance. A technical report published by the engi- 
neer Frank Hecker ("Setting up shop: the business of open source software", 
1998) [142] proposed that the best solution to the problem was to release the 
source code of the browser and benefit from the effects of the free software 
community, as described by Eric Raymond in "The Cathedral and the Bazaar". 
In January 1998, Netscape Inc. officially announced that it would publicly 
release the source code of its browser, marking an extremely important mile- 
stone within the short history of free software: a company was going to pub- 
lish the whole of the source code of an application that had been a commercial 
product up until then, under a free software license. The date of the launch 

was scheduled for the 31 st March 1998. 

GNUFDL» PID_001 48386 159 Free Software 

In the two months between January and March, the people at Netscape were 
frenetically active, trying to get everything ready The list of tasks was enor- 
mous and complicated ("Freeing the source: the story of Mozilla", 1999) [134]. 
On the technical level, it was necessary to contact the companies that made 
the modules to ask them for their consent to the change of license; if the 
answer was negative, the module had to be eliminated. In addition, all the 
parts written in Java had to be reimplemented, as it was considered that Java 
was not free. They then decided to call the free project Mozilla, just as the 
developers of Netscape had called their main component Mozilla, and the domain was purchased to build a community of developers and 
assistants based around this website. At the end of the process, more than one 
million and a half lines of source code were released. 


The name Mozilla is a play on words, with a little dose of humour from of the Netscape 
Inc. development team. The Mozilla name came from adapting the name Godzilla, the 
monster that caused mayhem in Japanese horror films from the fifties, to make it sound 
like Mosaic Killer, as the new browser, with more advanced technology, was supposed to 
render Mosaic obsolete. 

On another note, there was the legal question. The free licenses existing at 
that time did not convince the Netscape executives, who could not see how 
this could be "compatible" with the commercial nature of a business. Netscape 
wanted a more flexible license, that would make it possible to reach agree- 
ments with third parties so as to include their code regardless of the license 
or whether other commercial developers were to contribute to it, so that they 
could defend their financial interests howsoever they chose. And although 
they had not initially planned to create a new license, they eventually reached 
the conclusion that this was the only way they could achieve what they want- 
ed. This is how the Netscape Public License (NPL) was created: a license that 
was based on the basic principles of free software licenses, but that also gave 
certain additional rights to Netscape Inc, which also made it a non-free li- 
cense, from the perspective of the Free Software Foundation. When the draft of 
the NPL was published for public discussion, the clause providing additional 
rights to Netscape was heavily criticised. Netscape Inc. reacted quickly in re- 
sponse to these criticisms and created an additional license, the Mozilla Public 
License (MPL), which was identical to the NPL, except in that Netscape had 
no additional rights. 

The final decision was to release the Netscape code under the NPL license, 
which provided additional rights to Netscape, and that any new code that 
was included would be issued under the MPL (or a compatible license). The 
corrections to the original code (licensed with the NPL) would also be covered 
by this license. 

GNUFDL» PID_001 48386 160 Free Software 


Currently, Mozilla accepts contributions under three licenses: the MPL, the GPL and the 
LGPL. Changing the license was not at all easy, as they had to find all the people that had 
contributed code at any point so that they would give their consent to the changeover 
from the NPL/MPL to the MPL/GPL/LGPL. In order to relicense the whole code, a website, 
which contained a list of three hundred "lost" hackers, was created ("Have you seen these 
hackers?") [38]. As at May 2007, they are still looking for two of these developers. 

Developing the original code of Netscape Communicator was, without a 
doubt, more complicated than initially expected. The initial conditions were 
already bad to start with, because what was released was, on occasions, in- 
complete (all the third party modules for which no consent had been given 
for the release had been removed) and it hardly worked. As if that were not 
enough, apart from the technical problem of making Mozilla work on a large 
number of operating systems and platforms, there were the flaws taken from 
Netscape Inc., with release cycles that were too long and inefficient for the 
world of Internet and that did not distinguish between its own interests and 
the community formed around Mozilla. All of this came to a head exactly a 
year later when one of the most active programmers from before and after the 
release, Jamie Zawinsky, decided to throw in the towel in a bitter letter ("Res- 
ignation and post-mortem", 1999) [237] in which he made clear his despair 
and desolation. 

On 15 th July 2003, Netscape Inc. (now the property of America On Line) an- 
nounced that it was no longer going to develop the Netscape browser and, 
therefore, was no longer going to actively take care of the Mozilla project. 
As a kind of "redundancy settlement" Netscape approved the creation of the 
Mozilla Foundation, which it supported with a contribution of two million 
dollars. Likewise, all of the code that was under the NPL (Netscape's public 
license) was donated to the Foundation and redistributed with the licenses 
previously published by the Mozilla project: MPL, LGPL and GPL. 

On 10* March 2005, the Mozilla Foundation announced that it would not 
publish any more official versions of the Mozilla Application Suite, which 
would be replaced by Mozilla SeaMonkey, that included a web browser, an 
email client, an address book, an HTML editor and an IRC client. On another 
note, the Mozilla project hosts various independent applications, the most 
notable of which include Mozilla Firefox (web browser), which is undoubted- 
ly the most well-known, Mozilla Thunderbird (email and news client), Mozil- 
la Sunbird (calendar), Mozilla Nvu (HTML editor), Camino (web browser de- 
signed for Mac OS X) and Bugzilla (web-based bug-tracker tool). 

As time has progressed, despite the many doubts and the long periods in which 
it seemed that it was destined to fail, the project now seems to be going well. 
Thanks to the versatility and portability of its applications, despite requiring 
many runtime resources in many cases, they are used (generally, but especially 
Firefox) as the pair in the end user's desktop. 

GNUFDL' PID 00148386 


Free Software 

9.6.2. X-ray picture of Mozilla 

The figures that we will discuss in this section correspond to a study of Firefox, 
the most well-known of the project's applications. According to the estimates 
of the COCOMO model, a company that wished to create software of this 
scale would have to invest approximately 111 million dollars to obtain it. The 
time it would take would be about seven years and the average number of 
programmers working full-time that the company would have to use would 
be approximately one hundred and twenty. 

Table 14. Current status of Mozilla Firefox 


Beginning of the project 






Source code lines. 


Cost estimate 

$ 111,161,078 

Runtime estimate 

6.87 years (82.39 months) 

Estimate of average number of develc 



Approximate number of developers 

50 committers 

Development assistance tools 

CVS, mailing lists, IRC, Bugzilla. 

C++ and C are the languages that are used the most, in that order of priori- 
ty. Perl is used and this is mainly due to the fact that the development assis- 
tance tools created by the Mozilla project, such as BugZilla or Tinderbox, are 
designed in this language. What is surprising is the large amount of code lines 
in an assembly language in an end user application. An inspection of the code 
in the repository shows that, in effect, there are quite a lot of files encoded 
in assembly language. 

Table 15. Programming languages used in Mozilla Firefox 

Programming language 

Code lines 










1 .26% 










GNUFDL' PID 00148386 


Free Software 

Programming language 

Code lines 











9.7. is one of the star applications in the current free software 
scene. It is a multiplatform office application suite that includes the key ap- 
plications in an office desktop environment, such as a word processor (Writ- 
er), a spreadsheet (Calc), a presentation program (Impress), a graphics editor 
(Draw), a tool for creating and editing mathematical formulae (Math) and, 
finally, an HTML language editor (included in Writer). The interface provid- 
ed by is homogeneous and intuitive, with an appearance and 
functionalities similar to those of other office applications, especially the one 
that is most widely used today, Microsoft Office. 

Written in C++, includes Java's API and has its own compo- 
nents for embedded systems, which makes it possible to include, for example, 
tables from a spreadsheet in the word processor in a very simple and intuitive 
way. One of its advantages is that it can handle a large amount of file formats, 
including those of Microsoft Office. Its native file formats, unlike those of 
Microsoft's office suite, are based on XML, which shows that they are clear- 
ly committed to versatility, the ease of transformation and transparency. Cur- 
rently, has been translated into more than twenty five lan- 
guages and it runs on Solaris (its native system), GNU/Linux and Windows. 
Versions for FreeBSD, IRIX and Mac OS X are expected in the not-too-distant 
future. took its definitive name (OpenOffice, as everybody knows it, 
plus the .org tag) after a court case, in which it was accused of trademark vio- 
lation by another company. 

9.7.1. History of 

In mid-1980s, the company StarDivision was founded in the Federal Repub- 
lic of Germany, with the principal aim of creating an office application suite: 
StarOffice. In summer 1999, SUN Microsystems decided to purchase the com- 
pany StarDivision and make a significant commitment to StarOffice, with the 
clear intention of wresting away part of the market share conquered by Mi- 
crosoft at that time. In June 2000, the company launched version 5.2 of StarOf- 
fice, which could be downloaded gratis from the Internet. 

GNUFDL» PID_001 48386 163 Free Software 

However, StarOffice's success was limited, as the market was already strongly 
dominated by Microsoft's office package. SUN decided to change its strategy 
and, as occurred with Netscape and the Mozilla project, decided to take advan- 
tage of free software to gain importance and implement its systems. Conse- 
quently, the future versions of StarOffice (a proprietary product of SUN) would 
be created using (a free product) as a source, respecting the 
application programming interfaces (API) and the file formats and serving as 
the standard implementation. 

9.7.2. Organisation of aims to have a decision-making structure in which all the 
members of the community feel like participants. Consequently, a system was 
devised so that the decision-making process would have the greatest consen- 
sus possible. The project is divided into a series of subprojects 
that are taken on by project members, the assistants and one single leader. Of 
course, the members of a project may work on more than one project, as can 
the leader. However, no one can lead more than one project at a time. The 
projects are divided into three categories: 

• Accepted projects. These can be technical or non-technical. The leaders 
of each accepted project have a vote when it comes to making global de- 

• Native-lang projects. These are all the internationalisation and localisa- 
tion projects of Currently, as we have mentioned, there 
are more than twenty five teams that are working on translating the applications to different languages and conventions. As a 
set, native-lang projects have one single vote on global decisions. 

• Incubating projects. These are the projects promoted by the community 
(generally, they are experimental or small). They may become accepted 
projects after a period of six months. In effect, the com- 
munity can guarantee that the accepted projects are based on a real inter- 
est, as the mortality rate of new projects in the world of free software is 
very high. In total, the incubating projects have one vote on the decisions 

9.7.3. X-ray picture of 

The office suite comprises approximately four million lines of 
source code distributed throughout forty five thousand files. 

GNUFDL' PID 00148386 


Free Software 

The COCOMO model estimates that the work required to build a "clone" of would have to be provided by one hundred and eighty pro- 
grammers working full-time for almost eight years. According to the COCO- 
MO estimates, the development cost would be approximately 215 million dol- 

The results discussed in this section were obtained from a study of the source 
code of stable version 2.1 of 

Table 16. Current status of 


Beginning of the project 

June 2000 (first free versions) 





Source code lines. 


Cost estimate 

$ 215,372,314 

Runtime estimate 

8.83 years (1 05.93 months) 

Estimate of average number of developers 


Approximate number of developers 

200 commiters 

Development assistance tools 

CVS, mailing lists 

Where the programming languages used in are concerned, 
the most prevalent is C++. It is interesting to note how Sun's purchase of the 
company resulted in the integration of a lot of Java code in the office suite, 
which even exceeded the amount of language in C. 

Table 17. Programming languages used in 

Programming language 

Code lines 




88.81 % 









1 .04% 










GNUFDL» PID_001 48386 165 Free Software 

9.8. Red Hat Linux 

Red Hat Linux was one of the first commercial distributions of GNU/Linux. 
Today, it is probably one of the most well-known, and certainly the one that 
can be considered the "canonical" of all the commercial distributions. The 
work of the distributors is basically related to integration tasks and not so 
much to software development. Of course, Red Hat and other distributions 
may have developers working for them, but their work is secondary for the 
aims of a distribution. In general, it is assumed that the task of the distribu- 
tions is to simply take the source packages (generally the files published by 
the developers themselves) and bundle them so that they fulfil certain criteria 
(both technical and organisational). The product of this process is a distribu- 
tion: a series of properly organised bundles that make it possible for the user 
to install, uninstall and update them. 

Distributions are also responsible for the quality of the final product, which is 
a very important aspect if we consider that many of the applications that are 
included have been developed by volunteers in their free time. Consequently, 
the security and stability aspects are of the essence for a distribution. 

9.8.1. History of Red Hat 

Red Hat Software Inc. was founded by Bob Young and Marc Ewing in 1994. 
The main objective was to compile and commercialise a GNU/Linux distribu- 
tion that was called (and is still called) Red Hat Linux [236]. Basically, it was a 
bundled version of what existed on the Internet at that time, including docu- 
mentation and support. Version 1.0 of this distribution was born in the sum- 
mer of 1995. A few months later, in autumn, version 2.0, which included RPM 
technology (RPM package manager was published. The RPM package manager 
has become a de facto standards for packages in GNU/Linux systems. In 1998, 
version 5.2 of Red Hat was issued to the great public. For a complete history of 
the names of the different versions of Red Hat, please read "The truth behind 
Red Hat names" [201]. 


As of version 1 . 1 of Linux Standard Base (a specification designed to achieve binary com- 
patibility between GNU/Linux distributions, which is taken care of by the Free Standards 
Group), RPM has been chosen as the standard package manager. The Debian project con- 
tinues with its own package format, as do many distributions that depend on the Debian 
package management system, and they are adjusted to the standardised format using a 
conversion tool called alien. 

Before the RPM management system existed, almost all of the GNU/Linux dis- 
tributions offered the possibility of installing the software through a menu- 
based procedure, but making modifications to an existing installation, espe- 
cially adding new software packages after the installation, was not easy. RPM 
made that step beyond the state-of-the-art possible by providing users with the 
ability to manage their own packages ("Maximum RPM. Taking the Red Hat 

GNUFDL" PID_001 48386 166 Free Software 

package manager to the limit", 1998) [83], which made it possible to delete, 
install or update any software package existing in the distribution in a much 
easier way. The RPM package system continues to be the most widely used 
package management system in the different GNU/Linux distributions. The 
statistics of Linux Distributions, "Facts and figures", 2003 [92], a website that 
contains qualitative and quantitative information on a large number of dis- 
tributions, show that in May 2003, a large majority (sixty five) of the one 
hundred and eighteen distributions used for the calculations, used the RPM 
(approximately 55% of the total). In comparison, the Debian package format 
(known as deb) was only used in sixteen distributions (approximately 14% of 
the total). 

However, Red Hat Inc. was not only known for its software distribution based 
on Linux. In August 1999, Red Hat went public and its shares achieved the 
eighth highest first day gain in the whole of Wall Street history. Four years lat- 
er, the value of Red Hat's shares had shrunk to a hundredth of the maximum 
value they reached before the dotcom crisis. Nevertheless, its successful begin- 
nings on the stock market put Red Hat on the front pages of newspapers and 
magazines that did not specialise directly in IT matters. In any case, it seems 
that Red Hat has managed to overcome the problems that other companies in 
the business world have had with free software and the numbers it published 
in the last quarter of 2002 were in the black for the first time in its history. 

Another of the most important historical events involving Red Hat was the 
acquisition of Cygnus Solutions in November 1999, a company founded a 
decade before that had already proved how it was possible to earn money with 
an integral strategy based on free software ("Future of Cygnus Solutions. An 
entrepreneur's account") [216]. Cygnus chose the complicated market of com- 
pilers to make its mark. Its commercial strategy was based on the development 
and adaptation of GNU software development tools (basically GCC and GDB) 
tailored to the client's needs. 

In September 2003, Red Hat decided to concentrate its development work on 
the corporate version of its distribution and delegated the common version 
to Fedora Core, an open source project independent of Red Hat. 

In June 2006, Red Hat purchased the company JBoss, Inc., becoming the com- 
pany in charge of developing the most important open source applications 
server, J2EE. 

GNUFDL» PID_001 48386 167 Free Software 

9.8.2. Current status of Red Hat. 

Currently, Red Hat Inc.'s most important products are Fedora Core and Red 
Hat Network, an Internet software update service. These types of services are 
designed more with the end user in mind and not so much for the corporate 
environment, but they are good for Red Hat to advertise itself and to reinforce 
its brand strategy 

Red Hat's "real" commercial strategy is based on the products it designs for the 
corporate world. These types of products are much less well-known, but they 
constitute a major part of Red Hat's turnover, much greater than that of its 
most popular star products in the literal sense. 

Red Hat has a distribution that is corporate-orientated, integrated around an 
applications server called Red Hat Enterprise Linux AS. Clients that purchase 
this software also receive support. The equivalent of Red Hat Network for com- 
mercial users is Red Hat Enterprise Network, which includes system manage- 
ment and the option of obtaining updates. On anther note, Red Hat also offers 
IT consultancy services and a certification program similar to that offered by 
Microsoft in the world of Windows. 

9.8.3. X-ray of Red Hat 

Red Hat has recently passed the milestone of fifty million lines of code, which 
makes it one of the biggest software distributions that have ever existed, ex- 
ceeding, as we shall see later in this chapter, the size of proprietary operating 
systems. Red Hat Version 8.1 consisted of 792 packages, so we can assume that 
the latest version would have had more than eight hundred packages, if we 
consider that the number tends to increase slightly from version to version. 

As in our previous examples, the COCOMO model has been used to estimate 
the investment and effort that would have been necessary in order to create 
a generation of software of the same scale. However, in Red Hat's case, we 
have taken into account the fact that it is a product prepared using a series of 
independent applications. Consequently, an independent COCOMO estimate 
has been used for each one of Red Hat's packages, and then we have added 
the estimated costs and personnel that would have been required. In order 
to analyse the optimum design time for Red Hat, we have chosen the largest 
package, as, in theory, all the packages are independent and could therefore 
be designed at the same time. For this reason, the optimum design time for 
Red Hat is similar to that of the other projects presented in earlier sections 
of this chapter. 

GNUFDL' PID 00148386 


Free Software 

According to COCOMO, approximately seven and a half years and a team 
of developers, consisting of an average of one thousand eight hundred devel- 
opers, would have been necessary in order to design the Red Hat Linux 8.1 
distribution, starting from zero. The cost of the final development would be 
approximately 1,800 million dollars. 


One thousand eight hundred million dollars is the sum that the Spanish Ministry of 
Defence has allocated to renewing its helicopter fleet in the latest budget. Out of that 
sum, half will be invested in buying twenty four helicopters, so we could say that the 
price of Red Hat would be equivalent to that of forty eight combat helicopters. Likewise, 
1,800 million dollars is the total global earning from the film Titanic. 

Table 18. Status of Red Hat Linux. 


Beginning of the project 





Source code lines. 

More than 50,000,000 

Number of packages 


Cost estimate 

$ 1,800,000,000 

Runtime estimate 

7.35 years (88.25 months) 

Estimate of average number of develop- 


Approximate number of developers 

Red Hat employees (generally integration only) 

Development assistance tools 

CVS, mailing lists 

Due to the fact that there are many packages, the languages in Red Hat are 
more diverse than the ones we have seen in the most important free software 
applications. In general terms, C is very important, with more than sixty per 
cent of the code lines. In second place, with more than ten million lines of 
code, we have C++, followed by a long distance by Shell. It is interesting to 
note that after Perl we have Lisp (mainly due to its use in Emacs), assembly 
language (of which a quarter corresponds to the language that comes with 
Linux) and a language whose use is frankly declining, Fortran. 

Table 19. Programming languages used in Red Hat. 

Programming language 

Code lines 








GNUFDL' PID 00148386 


Free Software 




Code lines 










1 .92% 



1 .29% 



1 .07% 

9.9. Debian GNU/Linux 

Debian is a free software operating system that currently uses the Linux kernel 
for its distribution (although it is expected that there will be Debian distribu- 
tions based on other kernels in the future, as is the case of "the HURD"). It 
is currently (in 2007) available for various different architectures, including 
Intel x86, ARM, Motorola, 680x0, PowerPC, Alpha and SPARC. 

Debian is not only the largest existing GNU/Linux distribution, but also one 
of the most stable and it has received various awards for the fact that it is pre- 
ferred by users. Although its user base is difficult to estimate, as the Debian 
project does not sell CDs or any other media with its software and the software 
that it does have can be redistributed by anyone that wishes to, we can sup- 
pose, with a reasonable degree of certainty, that it is an important distribution 
within the GNU/Linux market. 

There is a categorisation in Debian that depends on the license and distribu- 
tion requirements of the packages. The kernel of the Debian distribution (the 
section called main that covers a great variety of packages) consists only of 
free software in accordance with the DFSG (Debian Free Software Guidelines) 
[104]. It can be downloaded from the Internet and many redistributors sell it 
on CDs or in other media. 

Debian distributions are created by almost one thousand volunteers (general- 
ly IT professionals and experts). The work of these volunteers consists of tak- 
ing the source programs, in most cases from the original authors, configuring 
them, compiling them and bundling them so that an average Debian distri- 
bution user only has to select the package and the system will install it with 
no further problems. What may first appear as simple can become complex as 
soon as other factors, such as the dependencies between the different packages 
(package A needs package B in order to work) and the different versions of all 
these packages, are taken into account. 

The work performed by the members of the Debian project is the same as that 
performed in any other distribution: the integration of the software so that it 
all works together properly. Apart from the adaptation and packaging work, 
the Debian developers are in charge of maintaining an Internet-based services 

GNUFDL» PID_001 48386 170 Free Software 

infrastructure (website, online files, bug management system, assistance mail- 
ing lists, support and development, etc.), various translation and internation- 
alisation projects, the development of various tools specific to Debian and, 
generally, in charge of anything that is required in order to make the Debian 
distribution work. 

Apart from its voluntary nature, the Debian project has a feature 
that is especially unique: Debian's social contract ( 
social_contract.html) [106] . This document does not only describe the Debian 
project's main goals, but also the means that will be used to achieve them. 

Debian is also known for having a very strict packages and versions policy, 
designed to achieve the best quality in the product (the "Debian policy man- 
ual") [105]. In this way, there are three different types of Debian at any given 
time: a stable version, an unstable version and a test version. As the name itself 
indicates, the stable version is the one recommended for systems and users 
that require complete stability. The software has to be subjected to a freeze 
period, during which any bugs are corrected. The general rule is that the stable 
Debian version must not have any known critical bug. On the other hand, 
this stable version does not usually have the latest versions of the software 
(the newest additions). 

There are another two versions of Debian that exist alongside the stable one 
for those that wish to have the most recent software. The unstable version 
includes packages that are being stabilised, whereas the test version, as the 
name indicates, is the one that has a greater tendency to fail and that contains 
the newest of the new in terms of the latest software. 

When the first study was performed, the stable version of Debian was De- 
bian 3.0 (also known as Woody), the unstable one was nicknamed Sid and 
the test version was Sarge. However, Woody also went through an unstable 
stage and, before that, a test stage. This is important, because what we will 
consider in this article will comprise the different stable versions of Debian, 
ever since version 2.0 was published, in 1998. For example, we have Debian 
2.0 (alias Hamm), Debian 2.1 (Slink), Debian 2.2 (Potato) and, finally, Debian 
3.0 (Woody). 


The nicknames of the Debian versions correspond to the main characters in the animated 
film Toy story, a tradition that started, half-jokingly and half-seriously, when version 2.0 
was published and Bruce Perens, the leader of the project at the time and subsequent 
founder of the Open Source Initiative and the phrase open source, was working for the 
company that was designing the film. For more details regarding Debian's history and 
the Debian distribution in general, we recommend "A brief history of Debian" [122]. 

GNUFDL' PID 00148386 


Free Software 

9.9.1. X-ray picture of Debian 

Debian GNU/Linux is probably the largest compilation of free software that 
works in a coordinated manner and, doubtlessly, one of the biggest software 
products ever built. Version 4.0, released in April 2007 (called Etch), consists 
of more than ten thousand source packages, with more than 288 million lines 
of code. 

The number of lines of code in Debian 3.0 is 105 million. According to the 
COCOMO model, a sum of approximately 3,600 million dollars would have 
to be paid in order to obtain software similar to that bundled with this distri- 
bution. As with Red Hat, the effort required to build each package separately 
has been calculated and the resulting figures have then been added to each 
other. For the same reason, the time it would have taken to develop Debian is 
only seven years, as the packages could have all been built at the same time as 
each other. However, an average of approximately four thousand developers 
would have had to have been mobilised during those seven years. 


Three thousand six hundred 
million dollars is the budget al- 
located by the 6 th EC Frame- 
work Programme for research 
and development on the in- 
formation society. It is also the 
sum that Telefonica intends to 
invest in Germany in order to 
deploy UMTS services. 

Table 20. Status of Debian 


Beginning of the project 



Those that fulfil the DFSG 

Version used 

Debian 4.0 (alias Etch) 

Source code lines. 


Number of packages 


Cost estimate 

$ 10,140 million 

Runtime estimate 

8.84 years 

Approximate number of maintainers 

Approximately 1,500 

Development assistance tools 

Mailing lists, bug report system 

The most commonly used language in Debian 4.0 is C, with more than 51% 
of the lines of code. However, as we shall show a little later in this section, the 
importance of C is declining with time, as 80% of the code in the first versions 
of Debian, was in C. The second most commonly used language, C++, shares 
a fair part of the "blame" for the decline of C; however, C has especially been 
influenced by the rise of scripting languages such as Perl, Python and PHP. On 
the other hand, languages such as Lisp or Java (which is underrepresented in 
Debian due to its policy of not accepting code that depends on Sun's propri- 
etary virtual machine) sometimes manage to get in. 

GNUFDL* PID 00148386 


Free Software 

Table 21. Programming languages used in Debian GNU /Linux 4.0 

Programming language 

Lines of code (in millions) 

























1 .24% 

Table 22 shows how the most important languages developed in Debian. 

Table 22. Languages most used in Debian 


Debian 2.0 

Debian 2.1 

Debian 2.2 

Debian 3.0 

















1 3,000,000 






















1 .68% 









1 .96% 


1 .98% 


1 .99% 


1 .84% 












1 .23% 


1 .24% 




1 .02% 

There are languages that we could consider to be in the minority that reach 
fairly high positions in the classification. This is due to the fact that, whilst 
they are only present in a small number of packages, the packages in question 
are quite big. Such is the case of Ada, which whilst only being in three pack- 
ages (GNAT, an Ada compiler; libgtkada, a link to the GTK library and ASIS, 
a system for managing Ada sources) covers 430,000 of the total 576,000 lines 
of source code that have been counted in Debian 3.0 for Ada. Another similar 
case is Lisp, which only appears in GNU Emacs and XEmacs, but has more than 
1,200,000 lines of the approximately four million in the whole distribution. 

GNUFDL* PID 00148386 


Free Software 

9.9.2. Comparison with other operating systems 

There is the proverb that says that all comparisons are odious; this is especially 
true when comparing free software with proprietary software. The detailed 
x-ray pictures taken of Red Hat Linux and Debian were possible because they 
are examples of free software. Having access to the code (and to the other 
information that has been provided in this chapter) is essential for studying 
the different versions' number of lines, packages, programming languages, etc. 
But the advantages of free software go beyond this, because, in addition, they 
make it easier for third parties, whether they are research teams or simply 
people that are interested, to analyse them. 

In proprietary systems in general, a study such as this would be completely im- 
possible. In fact, the figures provided below were obtained from the companies 
behind proprietary software development themselves, which means that we 
are not in a position to guarantee their truthfulness. To top this off, in many 
cases we do not know whether they are talking about physical source code 
lines, as we have done during this chapter, or whether they also include the 
blank lines and comments. Furthermore, neither do we know for certain what 
they include in their software, which means that we do not know whether 
certain versions of Microsoft Windows include the Microsoft Office suite or 

In any case, considering all that we have discussed on this matter in previous 
paragraphs, we believe that it is interesting to include this comparison, as 
it helps us to see the position in which the different Red Hat and Debian 
distributions are in, within a wider context. What is unquestionable is that 
both Debian and Red Hat, but especially the former, are the largest collections 
of software ever seen by humanity to date. 

The figures cited below come from Mark Lucovsky [168] for Windows 2000, 
SUN Microsystems [171] for StarOffice 5.2, Gary McGraw [169] for Windows 
XP and Bruce Schneier [200] for all the other systems. Table 23 provides a 
comparison, from smallest to greatest. 

Table 23. Comparison with proprietary systems 


Date of 


Lines of code 


Microsoft Windows 3.1 

April 1992 


SUN Solaris 7 

October 1 998 


SUN StarOffice 5.2 

June 2000 


Microsoft Windows 95 

August 1995 


Red Hat Linux 6.2 

March 2000 


Debian 2.0 

July 1998 


GNUFDL* PID 00148386 


Free Software 


Date of publication 

Lines of code (approx.) 

Microsoft Windows 2000 

February 2000 


Red Hat Linux 7.1 

April 2001 


Debian 2.1 

March 1999 


Windows NT 4.0 

July 1996 


Red Hat Linux 8.0 

September 2002 


Debian 2.2 

August 2000 


Debian 3.0 

July 2002 


9.10. Eclipse 

The Eclipse platform consists of an open and extensible IDE (integrated devel- 
opment environment). An IDE is a program consisting of a set of tools that are 
useful for a software developer. The basic elements of an IDE include a code 
editor, a compiler/interpreter and a debugger. Eclipse is an IDE in Java and 
provides numerous software development tools. It also supports other pro- 
gramming languages, such as C/C++, Cobol, Fortran, PHP or Python. Plug-ins 
can be added to the basic platform of Eclipse to increase the functionality. 

The term Eclipse also refers to the free software community that develops the 
Eclipse platform. This work is divided into projects with the aim of providing 
a robust, scalable and quality platform for the development of software with 
the Eclipse IDE. The work is coordinated by the Eclipse Foundation, which 
is a non-profit organisation created for the promotion and development of 
the Eclipse platform and that supports both the community and the Eclipse 

9.10.1. History of Eclipse 

A lot of Eclipse's programming was carried out by IBM before the Eclipse 
project was created as such. Eclipse's predecessor was VisualAge and it was 
built using Smalltalk in a development environment called Envy. After Java 
appeared in the nineties, IBM developed a virtual machine that worked with 
both Smalltalk and Java. The rapid growth of Java and its advantages with 
the focus on an Internet that was expanding heavily forced IBM to consider 
abandoning this dual virtual machine and to build a new platform based on 
Java from scratch. The final product was Eclipse, which had already cost IBM 
approximately 40 million dollars in 2001. 

Towards the end of 2001, IBM, along with Borland, created the non-profit 
Eclipse foundation, thereby opening up to the open source world. This consor- 
tium was gradually joined by important global software development compa- 
nies: Oracle, Rational Software, Red Hat, SuSE, HP, Serena, Ericsson and Nov- 

GNUFDL" PID_001 48386 175 Free Software 

ell, among others. There are two significant absences: Microsoft and Sun Mi- 
crosystems. Microsoft was excluded due to its monopoly of the market and 
Sun Microsystems had its own IDE, constituting Eclipse's main competition: 
NetBeans. In fact, the Eclipse name was chosen because the aim was to create 
an IDE able to "eclipse Visual Studio" (Microsoft) and to "eclipse the sun" (Sun 

The latest stable version of Eclipse is available for the Windows, Linux, Solaris, 
AIX, HP-UX and Mac OS X operating systems. All versions of Eclipse need to 
have a Java Virtual Machine QVM) installed in the system, preferably JRE (Java 
Runtime Environment) or JDK (Java Developer Kit) by Sun, which, as at early 
2007, are not yet free (although Sun has announced that their JVM will be). 

9.10.2. Current state of Eclipse 

All the work prepared for the Eclipse consortium is organised into different 
projects. These projects are in turn divided into subprojects and the subpro- 
jects into components. The high-level projects are managed by committees of 
the Eclipse Foundation (PMC, project management committees). The following 
list shows the high-level projects: 

• Eclipse. Base platform for the rest of the components. This platform will 
be free, robust, complete and of a good quality for the development of rich 
client platforms (RCP) and integrated tools (plug-ins). The Eclipse platform's 
runtime kernel is called Equinox and it is an implementation of the OS- 
Gi specification (Open Services Gateway Initiative), which describes a ser- 
vices oriented architecture (SOA) for applications. 

• Tools (ETP, Eclipse tools project). Various tools and common components 
for the Eclipse platform. 

• Web (WTP, web tools project). Tools for the development of web applications 
and JEE (Java Enterprise Edition). 

• Test and performance tools project (TPTP). Testing tools and performance 
level measurers so that the developers can monitor their applications and 
make them more productive. 

• Web reports (BIRT, business intelligence and reporting tools). Web report gen- 
eration system. 

• Modelling (EMP, Eclipse modelling project). Model-based development tools. 

• Data (DTP, data tools platform). Support for data-handling technologies. 

GNUFDL» PID_001 48386 176 Free Software 

• Embedded devices (DSDP, device software development platform). Tools for 
the development of applications that are to be run on devices with limited 
hardware, in other words, embedded devices. 

• Service oriented architecture (SOA). Tools for developing service-oriented 

• Eclipse Technology. Research, dissemination and development of the 
Eclipse platform. 

The principles that guide the development of the Eclipse community are as 

• Quality. The software developed at Eclipse must meet the software engi- 
neering quality standards. 

• Development. The Eclipse platform, and all the tools based on it, must 
develop dynamically in accordance with the users' requirements. 

• Meritocracy. The more someone contributes, the more responsibilities he 
or she has. 

• Eclipse Ecosystem. There will be resources donated by the open source 
community to the Eclipse consortium. These resources will be employed 
in ways that benefit the community. 

Eclipse's development process follows certain predefined phases. Firstly, there 
is a phase called the pre-proposal phase, in which an individual or company 
declares their interest in establishing a project. If the proposal is accepted, it 
is decided whether it will be a high-level project or a subproject. The next 
step is to validate the project in terms of applicability and quality. After a 
phase in which the project is incubated, there will be a final revision. If the 
project passes this revision, it will have proved its validity before the Eclipse 
community and it will pass into the implementation phase. 

9.10.3. X-ray of Eclipse 

Eclipse is distributed under an EPL License (Eclipse Public License). This license 
is considered free by the FSF and the OSI. Under the EPL License, it is possible 
to use, modify, copy and distribute new versions of the licensed product. EPL's 
predecessor is the CPL (Common Public License). The CPL was written by IBM, 
whereas the EPL is the work of the Eclipse consortium. 

Estimating the investment and effort put into Eclipse is not an easy task. This 
is due to the fact that the source code that comprises the Eclipse ecosystem is 
distributed in numerous projects and software repositories. 

GNUFDL* PID 00148386 


Free Software 

Below are the results of applying the COCOMO model to the Eclipse platform, 
which is used as the base for the rest of the plug-ins. 

Table 24. Analysis of Eclipse 


Beginning of the project 



Eclipse Public License 

Analysed version 


Source code lines. 


Number of files 


Cost estimate 

$ 85,831,641 

Runtime estimate 

6.22 years (74.68 months) 

Estimate of average number of devel- 


Approximate number of developers 

1 33 commiters 

Development assistance tools 

CVS, mailing lists, bug-tracking system (Bugzilla) 

The following table shows the programming languages used in Eclipse 3.2.2: 

Table 25. Programming languages used in Eclipse 

Programming language 

Code lines 
































GNUFDL" PID_001 48386 178 Free Software 

10. Other free resources 

"If you want to make an apple pie from scratch, you must first create the universe." 
Carl Sagan 

Can the ideas behind free programs be extended to other resources? We could 
consider that other information resources that can easily be copied electroni- 
cally are similar to programs and that the same freedoms, rules, development 
and business models could apply to them. However there are some differences 
and the implications of these differences have meant that they have not de- 
veloped with the same force as programs. The main difference is that all one 
has to do is copy the programs to make them work, whereas when other types 
of information are copied, they have to pass through a more or less costly 
process before they can begin to be useful in any way, which can go from 
learning a document to the production phase of hardware described in the 
appropriate language. 

10.1. The most important free resources 

We already discussed the documentation of programs and other technical doc- 
uments in section 3.2.5. Here we will look at other types of creations, which 
can also be textual, but which are not related to software, but rather to scien- 
tific, technical and artistic fields. 

10.1.1. Scientific papers 

The way in which science evolves is, to a large extent, due to the fact that 
the researchers that make it progress for the benefit of humanity publish the 
results of their work in journals that reach a wide public. Thanks to this dis- 
semination, researchers develop a track record that allows them to progress 
towards positions of higher standing and responsibility, whilst they receive 
income from research contracts that they obtain thanks to their developing 

This way of disseminating papers represents a business model that has proved 
very fruitful. For this model to work, the quality of the work has to be guaran- 
teed and the papers must be widely disseminated. The obstacle that prevents 
the dissemination is the large amount of existing journals, of a significant 
cost, which can only be purchased with generous budgets. The quality is guar- 
anteed by the fact that the papers are reviewed by specialists or peers. 

In relation to this, numerous online journals have emerged, among which we 
would mention the veteran First Monday ("First Monday: peer reviewed jour- 
nal on the Internet") [26] or the Public Library Of Science project (PLOS http:// 

GNUFDL' PID 00148386 


Free Software [55]). The "Directory of Open Access Journals" 
[22] cites many more. Should persons other than the authors be allowed to 
publish modifications to these types of papers? There are objections that range 
from the possibility of substandard quality or equivocation of opinions or re- 
sults, to the danger of people that can easily plagiarise the papers and rise in 
the ranks with no effort, whilst denying the true authors of their hard-earned 
merits. However, the fact that all writers are under the obligation of citing the 
original author and of submitting the paper to a peer-review for publication 
in a prestigious journal can offset these problems (see section 10.2.2). 

An analogy has been established between free software and science, as the de- 
velopment model of the former requires the greatest amount of dissemination, 
peer-reviews (presumably experts) and the reuse of results ("Free software/free 
science", 2001) [154]. 

10.1.2. laws and standards. 

There are documents of a regulatory nature that define how things must be 
done, so as to improve coexistence between people or so that programs or 
machines can operate together. These documents need to be widely dissem- 
inated, which means that any obstacles will be counterproductive. For this 
reason, it is understandable that they receive special treatment, as exemplified 
in the Spanish Intellectual Property Act: 

"Legal or regulatory provisions and drafts thereof, judgements of jurisdictional bodies 
and acts, agreements, deliberations and rulings of public bodies, and official translations 
of all such texts, shall not be the subject of intellectual property". 

The technological equivalent of these laws would be the norms or standards. 
In programming, the communications protocols, either between remote ma- 
chines or between modules in the same machine, are especially important. It 
is obvious that we should not limit their dissemination, especially if we want 
free programs that operate with others to flourish, but, despite this, tradition- 
ally, the bodies that regulate these matters, such as ISO 11 and ITU 12 , sell their 
regulations and standards, even in electronic formats, and prohibit their re- 
distribution. Although this can be justified to an extent, claiming the need 
to cover part of the costs, the free dissemination of the standards has been 
much more productive; this is the case of the W3C 13 guidelines and, especially 
where Internet standards are concerned, the documents called RFCs (request 
for comments) that have existed since the beginning, in electronic formats that 
can be read using any form of text editor. 

( 'International Organisation for 

( International Telecommunica- 
tions Union 


World Wide Web Consortium 

However, the success of the Internet protocols is not due solely to their avail- 
ability. Other factors include the development model, which is very similar to 
free software due to its openness to the participation of any interested person 

GNUFDL» PID_001 48386 180 Free Software 

and the use of mailing lists and similar elements. This process is described 
in "The Internet standards process - revision 3" [94] and "The Tao of IETF: A 
Novice's Guide to the Internet Engineering Task Force" [136]. 

Should modifying the texts of laws and regulations be allowed? Obviously not 
if it leads to confusion. For example, an RFC should only be modified in order 
to explain it or add clarifying comments, whereas not even this is allowed 
without an explicit authorisation for the recommendations of the W3C (http:/ 
/www. [65]. 
The licenses themselves are also legal documents that cannot be modified. 
Should it be possible to create new regulations derived from other existing 
ones using the original documents? This would probably lead to the effortless 
spread of similar and incompatible regulations that would create confusion 
and could help the companies that dominate the market to promote their 
own incompatible variations, as it is in fact occurring, especially in the sphere 
of the Internet. Nevertheless, where State legislation is concerned, very often 
the laws have been copied literally from those of other countries and adapted 
with small modifications to the local particularities. 

Is there a business model for laws and regulations? There are numerous pro- 
fessionals that work on the laws, in charge of designing, interpreting and en- 
forcing them (legislators, lawyers, solicitors, judges, etc). There are laboratories 
that provide compliance certificates for the regulations. The regulatory bodies 
subsist, or should subsist, on the contributions of their members who wish to 
promote standards, for example, because their business is based on products 
that interoperate. 

In the same way that it is convenient to have a definition of free software or 
open software, it is also necessary to have a working definition of open standards. 
Bruce Perens ( [15] proposed the following 
definition based on the following principles: 

1) Availability: if possible, open standards must be available for all to read 
and implement. 

2) Maximise end user choice. 

3) Open standards must be free for all to implement with no royalty or fee 
(certifications of compliance may involve a fee, although Bruce Perens 
advises that there should be free self-certification tools available). 

4) No discrimination to favour one implementer over another. 

5) Extension or subset permissions (non-certifiable). 

GNUFDL» PID_001 48386 181 Free Software 

6) Avoidance of predatory practices by dominant manufacturers. All propri- 
etary extensions must have an open standard implementation. 

10.1.3. Encyclopaedias 

In 1999, Richard Stallman proposed the idea of a free encyclopaedia ("The free 
universal encyclopaedia and learning resource", 2001) [210] as a mechanism 
for avoiding the appropriation of knowledge and providing universal access to 
learning and the associated documents. It would consist of articles provided 
by the community, with no centralised control, where different actors would 
undertake different tasks, including, as a recommendation but not an obliga- 
tion, that of revising or checking the articles. This encyclopaedia would not 
only contain text but also multimedia and free educational software. 

Various initiatives have emerged to make this a reality. For instance, Nupedia 
( [178] tried to build a quality encyclopaedia, but 
the attempt failed, perhaps because it required a format that was relatively 
difficult to learn (TEI), although probably more because of the requirement of 
having all the articles edited, revised by scientists and checked for style, etc. 

The successor to Nupedia, which was much more successful, was Wikipedia 
( [69]. Wikipedia is a free multilingual encyclopae- 
dia based on wiki technology. Wikipedia is written cooperatively by volun- 
teers and the vast majority of articles can be modified by anyone with a web 
browser. Its success is based on its structure, which is more flexible in terms of 
editing, which eliminates the obstacles that Nupedia had in place and which 
makes it closer to what Stallman had in mind. The word wiki comes from the 
Hawaiian wiki wiki ('quick'). Wiki technology allows anyone to edit any docu- 
ment using the structured text system, which is extraordinarily simple as we 
saw in section 8.6.2. In February 2007, the number of articles in English in 
Wikipedia was more than 1,500,000. 

GNUFDL» PID_001 48386 182 Free Software 


Wikipedia is a project by the non-profit organisation Wikimedia, which also has the 
following projects, based on the same model as Wikipedia: 

• Wiktionary ( [66]. This is a cooperative project that aims 
to create a free multilingual dictionary, with definitions, etymologies and pronunci- 
ations, in the required languages. 

• Wikibooks ( [67]. This is a project that aims to provide 
textbooks, manuals, tutorials or other pedagogic texts to anyone requiring these el- 
ements, for free. 

• Wikiquote ( [70], It is a compilation of famous phrases in 
all languages, which includes the sources when these are known. 

• Wikisource. It is a library of original texts that are in the public domain or that have 
been published with a GFDL (GNU free documentation license). 

• Wikispecies ( [71]. It is an open repertory of animal 
species, vegetable species, fungi, bacteria and all forms of known life. 

• Wikinews ( [68] . It is a source of free news content in which the 
users are the editors. 

• Commons ( [19]. It is a free repository of images 
and multimedia content. 

• Wikiversity ( [72]. It is an open and free educational platform, 
based on teaching projects at all educational levels. 

• Meta-Wiki ( [48]. It is the website that supports all the 
projects of the Wikimedia Foundation. 

We should also mention the Concise Encyclopedia of Mathematics, which has a 
more limited concept of what free means (it can only be consulted on Internet) 
and a development model in which it is necessary to submit all contributions 
to an editorial committee before publication. 

10.1.4. Courses 

With the same aim as the encyclopaedias, it is possible to produce free teach- 
ing materials, such as notes, slides, exercises, books, syllabic or didactic soft- 
ware. There is a tendency to view universities as businesses that produce and 
sell knowledge, which contradicts their basic principles. The reasons why a 
university may make these materials available to all are as follows: 

• The fulfilment of its mission, as an agent that disseminates knowledge. 

• The low cost of making existing materials available worldwide. 

• The fact that these materials cannot replace teaching in person. 

• The idea of these materials as publicity that may attract students and con- 
tribute to the university's prestige. 

GNUFDL» PID_001 48386 183 Free Software 

• The possibility of creating a community of teachers that review and im- 
prove the materials. 

The most prominent initiative in this area is that of the MIT (http:// [174], which has the aim of making more than two thousand 
well-catalogued resources accessible in a coherent and uniform manner. 

10.1.5. Collections and databases 

The mere compilation of information following determined criteria, organis- 
ing it and making it available is, in itself, a product of valuable information, re- 
gardless of the information itself, which is therefore the product of its authors 
and, consequently, subject to restrictions on the freedom to access, modify or 
redistribute the content. Therefore, if we want free information, we can also 
want free collections. 

For example, we may wish to classify important information in the Internet, 
organising and commenting the links. This is what the ODP (Open Directory 
Project [109]) does; it is operated by Netscape and is main- 
tained by voluntary editors organised according to a hierarchical structure. 
The full directory can be freely copied in RDF format and published with cer- 
tain modifications, as does Google and many other search engines that take 
advantage of it. Netscape, which owns the directory, guarantees an "Open Di- 
rectory Project social contract" [53] inspired on that of the Debian distribu- 
tion ( [106], which facilitates ex- 
ternal contributions ensuring that the Open Directory Project will always be 
free, with public policies, self-regulated by the community and the users as 
the first priority. 

Other examples of collections that might interest us are the free software dis- 
tributions, with the programs modified so that they fit together perfectly and 
are precompiled so that they can be run easily. 

10.1.6. Hardware 

There are two main aspects involved in freedom as regards to hardware. The 
first one is the need for the interfaces and instruction sets to be free, in such a 
way that anyone can create a device handler or a compiler for an architecture. 
The second point is that there should be sufficient information and power 
available for reproducing a hardware design, modifying it and combining it 
with others. The designs can be considered software in an appropriate lan- 
guage (VHDL, Verilog, etc). However, making them work is not easy, as they 
have to be manufactured, which is expensive and slow. However, there are 
initiatives in this sense, among which we could mention OpenCores (http:// [52], for integrated circuits. 

GNUFDL" PID_001 48386 184 Free Software 

10.1.7. Literature and art 

To finish off our examination of free resources, we cannot forget art and liter- 
ature, whose ultimate objective is not as much utilitarian as it is aesthetical. 
What reasons might an artist have to give people the freedom to copy, mod- 
ify or redistribute their work? On the one hand, it can help to make them 
well-known and favour the dissemination of their work, which allows them to 
obtain income from other activities, such as concerts or commissions, and on 
the other, it can promote experimentation and creativity. In art, we have the 
same circumstances as in technical subjects. Innovation is incremental and 
it is sometimes difficult to distinguish between plagiarism and a work that is 
representative or follows an artistic movement or trend. 

Obviously, creation and interpretation are not the same thing, and neither 
are music and literature. Music, painting, photography and cinema are very 
similar to programs, in the sense that they can be made to "work" immediately 
on a computer, whereas the same does not apply to sculpture, for example. 
There are not many open source initiatives in art and literature and the ones 
that exist are very diverse. We could mention the novels by the Wu Ming 
( [29] collective. 

10.2. Licenses for other free resources 

The licenses for free software have been a source of inspiration for other in- 
tellectual resources, in such a way that many of them have been adopted di- 
rectly, especially where documentation is concerned, and on other occasions, 
they have been adapted slightly, as occurs with the pioneering Open Audio Li- 
cense ( [114]. Most of these 
licenses are copyleft licenses, if they permit derived works. 

GNU's free documentation license (see section 10.2.1) has been used and is 
often used for all kinds of texts, although the Creative Commons licenses (see 
section 10.2.2) are gradually being accepted. 

In fact, program licenses (GPL and LGPL) have even been used for hardware, 
although this subject is complex and difficult to reconcile with the current 
law. In effect, the designs and diagrams can be used, without physically being 
copied, to extract ideas that are used for new closed designs. For example, the 
OpenlPCore Hardware General Public License ("OpenlPCore hardware general 
public license") [155] establishes that this appropriation is not permitted, but 
the legal validity of the document is questionable [209] . The only possible way 
of protecting these ideas is using some form of free patent, which is something 
that has not yet developed and is out of the reach of those that do not intend 
or are unable to establish a business built on the ideas. 

GNUFDL" PID_001 48386 185 Free Software 

10.2.1. GNU free documentation license 

One of the most well-known copyleft licenses for technical documentation, 
whether it corresponds to programs or any other matter, is that of the Free 
Software Foundation. After realising that a document is not the same as a 
program, Richard Stallman promoted a license for the documents that went 
with the programs and for other documents of a technical or didactic nature. 

In order to smooth the development of the derived versions, a transparent copy 
of the document must be made available to whoever needs it, as explained in 
section 3.2.5, as well as the opaque copies, in an analogy between the source 
codes and the objects of the programs. 

One of the reasons for having a license is to establish authorship and to ensure 
that the ideas or opinions expressed by the author are not mischaracterised. 
This is why the derived works must have a title on the cover different to that 
of the previous versions (unless express permission has been given) and must 
expressly state the place where the original can be obtained. The names of 
the main authors of the original documents must also be listed, as well as the 
names of the people that have made any of the modifications, and all notes on 
intellectual property must be preserved. Likewise, any acknowledgements and 
dedications must be preserved and the history section, if there is one, must be 
respected when new modifications are added. It is even possible, and this is 
the aspect of the license that has most been criticised, to designate invariant 
sections and cover texts, which no one can modify or eliminate, although the 
license only permits non-technical texts to be considered as invariant sections, 
which the license refers to as secondary sections. 

This license has created a lot of controversy in the free software world, to the 
point that the Debian distribution project is currently (at the time of publi- 
cation of this book) discussing whether to remove from debian the contents 
under this license or designate all documents that have the license as non-free 
and consider them as non-official. Even though there are no invariant sec- 
tions, because the derived works may be subject to the terms of the same li- 
cense, it is important to remember that they could be added subsequently. 
It is argued, for example, that there may be incorrect or obsolete invariant 
sections, which, nevertheless, have to be preserved. In any case, the license is 
incompatible with Debian's free software guidelines ( 
social_contract.html#guidelines) [104], but the question hinges perhaps on 
whether the documentation must follow these guidelines (for example, the 
texts of the licenses cannot be modified either). 


The first versions of this text were covered under the GFDL license, but the authors sub- 
sequently decided to use a Creative Commons license as well (see section 10.2.2), which 
is more appropriate for the characteristics of a book. So this text is a dual licensing work. 

GNUFDL* PID_001 48386 186 Free Software 

10.2.2. Creative Commons licenses 

Creative Commons ( [21] is a non-profit organi- 
sation that was founded in 2001 by experts in intellectual property and law in 
the information society, with the aim of fostering the creation, conservation 
and accessibility of intellectual resources ceded to the community in numer- 
ous ways. It is based on the idea that some people may not wish to make use 
of all the intellectual property rights that the law grants them, as this could 
impede their wide distribution. 

The first Creative Commons licenses for creative works, of which there were 
various versions, originally came about in late 2002. These licenses were de- 
signed to be: 

• strong enough to withstand a court's scrutiny, in many countries; 

• simple enough for non-lawyers to use; 

• sophisticated enough to be identified by various web applications. 

The different licenses allow the creator to select what types of freedoms are 
allowed, apart from copying, in accordance with four points: 


Attribution. The material can be distributed, copied or 
exhibited by third parties as long as the original author is credited. 

or ^^ Non-commercial. The original material and derivative 
works can be distributed, copied or exhibited for non-commercial use. 

No derivative works . The material can be distributed, copied or exhibited 
but may not be used to create new works deriving from the original. 

Share alike. The material can be modified and distributed but under the 
same license terms as the original material. 

In version 1.x of the Creative Commons licenses, there were eleven types of 
license, which combined the four basic characteristics mentioned above. 98% 
of the authors chose the "attribution" option; consequently, as of version 2.x 
of the Creative Commons licenses, attribution is a requirement. This reduces 
the eleven types of license to six, which are as follows: 

GNUFDL* PID 00148386 


Free Software 



[BY:lf = 

Attribution - No derivative works 





Attribution - No derivative works - Non-commercial 

Attribution - Non-commercial 

Attribution - Non-commercial - Share alike 

Attribution - Share alike 

The following table shows a schematic of the licenses with the corresponding 
icons. This icon is usually a link to a summary of the license, hosted at Creative 
Commons website[21]. 

commercial use. 

Does not allow 
commercial use. 



modifications if 

shared alike. 

Does not allow 


It is possible to use the generic icon instead of the icon representing the 
license, but it must be linked to the license chosen by the author. The HTML 
code of the link to the license may be obtained from Creative Commons [21]. 
Once the license has been chosen and the corresponding icon added, the work 
will have been licensed and you will receive the: 

• Commons deed. A summary of the license with the relevant icons for it. 
This summary will be shown when clicking on the link obtained from 
Creative Commons [21]. 

• Legal Code. This is the complete legal text on which the license is based. 
This text may be accessed from the summary mentioned above. 

• Digital code. This is the RDF (resource description framework) description, 
which search engines and other applications can use to identify the license 
and the terms of use. 



GNUFDL* PID_001 48386 188 Free Software 

In February 2007, version 3.0 of the Creative Commons licenses was pub- 
lished. This is an update that corrects many of the faults that people identified. 
The first large modification is that the generic license is no longer based on 
the US model and is now based on the terminology of the Berne Convention. 
Secondly moral rights and rights management societies are mentioned specif- 
ically as different rulings had been made in each jurisdiction. Thirdly and 
finally, the texts of both the commons deed and the legal code that went with 
each license were modified to make it clear that the clause on the recognition 
of authorship does not allow the licensee to imply or give the impression that 
they have a relationship or are associated in any way with the licensor. 

In addition, Creative Commons provides other types of licenses for specific 
applications. Such as: 

Public Domain. License used to release the work from copyright completely. 

Developing Nations. The most permissive license for countries considered to 
be in development by the World Bank. 

Sampling. License used for sharing snippets (fragments of code which 
perform a useful function). 



^J Founders' Copyright. License used to release the work from copyright after 
a period of fourteen or twenty-eight years. 

CC-GNU GPL. A license which adds the Creative Commons' metadata and 
summary (Commons Deed) to the Free Software Foundation's GNU 
General Public License. 

A license which adds the Creative Commons' metadata and summary 
(Commons Deed) to the Free Software Foundation's GNU General Public 


Wiki. License for Wiki. In practical terms this is identical to the attribution 
and share alike license. 

Music Sharing. License used to share music. 

Not all Creative Commons licenses are considered free by sectors linked to 
free software, as the four essential freedoms must apply before licenses are 
defined as such (see section 1.1.1) Benjamin "Mako" Hill (Debian and Ubuntu 
developer) created the ( [28] 
website, with the aim of providing a better definition of what is free culture 

GNUFDL» PID_001 48386 189 Free Software 

and what is not. On this basis, of the six basic Creative Commons licenses, 
only two are strictly free: attribution alone (BY) and attribution-share-alike 
(BY-SA), the latter of which also has copyleft. 

GNUFDL" PID_001 48386 191 Free Software 


[1] Aap Project: 

[2] Ada Core Technologies: 

[3] Alcove: 

[4] Alcove-Labs: 

[5] Alioth: 

[6] Anjuta: 

[7] The Apache Ant Project: 

[8] Arch Revision Control System: 

[9] artofcode LLC: 






Bazaar GPL Distributed Version Control Software: 

Berlios. The Open Source Mediator: 

Bitkeeper Source Management: 

Bruce Perens: 


Cisco Enterprise Print System: 



Concurrent Version System: 

Creative Commons, 

Directory of Open Access Journals: 

Eclipse - An Open Development Platform: 


eCos license 2.0: 

First Monday. Peer Rewiewed Journal on the Internet 

Free Software Foundation: 

Freedom Defined (Free Cultural Works): 

Fundacion Wu Ming: 



GNU Automake: 

GNU Emacs: 

GNU Libc: 

GNU Libtool: 

GNUFDL» PID 00148386 192 Free Software 



GNU Make: 

GNU Troff: 

"Have you seen these hackers?": 

"History of TeX": 

IBM Public License Version 1.0: 

Jam Product Information: 



The Linux Documentation Project: 


Mailman, the GNU Mailing List Manager: 

The Malone Bug Tracker: 


Mozilla Public License 1.1: http://www.m0zilla.0rg/MPL/MPL-l.l.html 

Mozilla Tinderbox: 


Open Cores: 

Open Directory Project Social Contract: 

Open Source Initiative: 

Public Library of Science: 

Red Hat: 

Savannah: and 

Slashdot: News for Nerds, 

Sleepycat License: 

Sleepycat Software: 

SourceForge: Open Source Software Development Website: 


Texinfo - The GNU Documentation System: Open Source Software Engineering: 

W3c Document License: 


GNUFDL» PID_001 48386 193 Free Software 

[71] Wikispecies: 

[72] Wikiversity: 

[73] X Window System Release 11 License: 

[74] Ximian: 

[75] Zope Corporation: 

[76] Zope Public License 2.0: 

[77] Law on Intellectual Property. Spanish Royal Legislative Decree 1/1996, of 12 th April (April 

[78] Affero General Public License, 2002: 

[79] Law on Intellectual Property. Spanish Law 23/2006, of 7 th July (July 2006): 

[80] Flossimpact Study. Technical Report, European Comission, 2007: 

[81] ISO JTC 1/SC 34. Standard Generalised Markup Language (SGML, ISO 8879), 1986: 

[82] Antoniades, I.; Samoladas, I.; Stamelos, I.; Bleris, G. L. "Dynamical simulation 
models of the open source development process" En: Koch [157]. 

[83] Bailey, E. C. (1998). Maximum RPM. Taking the Red Hat package manager to the limit. 

[84] Gonzalez Barahona, J. M. (2000). "Software libre, monopolios y otras yerbas". Todo 
Linux (3). 

[85] Gonzalez Barahona, J. M. (2002). "iQue se hace con mi dinero?". Todo Linux (17). 

[86] Gonzalez Barahona, J. M.; Robles, G. Libre Software Engineering Web Site. 

[87] Gonzalez Barahona, J. M.; Robles, G. (2003, mayo). "Unmounting the code god 
assumption". En: Proceedings of the Fourth International Conference on extreme Programming and 
Agile Processes in Software Engineering. Genoa, Italy. 

[88] Gonzalez Barahona, J. M.; Robles, G.; Ortuno Perez, M. A.; Rodero Merino, 
L.; Centeno Gonzalez, J.; Matellan Olivera, V.; Castro Barbero, £. M.; De las Heras 
Quiros; P. "Anatomy of two GNU/Linux distributions". En: Koch [157]. 

[89] Barnson, M. P. The Bugzilla guide. 

[90] Baudis, P. "Cogito manual page". 

[91] Bezroukov, N. (1998, diciembre). "A second look at the cathedral and the bazaar". First 
Monday, 4(12). 

[92] Bodnar, L. (2003). "Linux distributions. Facts and figures". 

http://www. php?section=packagemanagement 

[93] Boehm, B. W. (1981). Software Engineering Economics. Prentice Hall. 

GNUFDL» PID_001 48386 194 Free Software 

[94] Bradner, S. (1996, October). "The Internet standards process. Revision 3 (rfc 2026, bcp 

[95] Cederqvist, P.; GNU (1993). "CVS - concurrent versions system", 

[96] Collins-Sussman, B.; Fitzpatrick; B. W.; Pilato, C. M. (2004). Version control with 
Subversion. O'Reilly & Associates ( 

[97] Cunningham, W. "Wiki design principles". 

[98] Dachary, L. (2001). "Savannah, the next generation". 

[99] Autonomous Government of Andalucia (2003, March). Decree 72/2003, of 18 th 
March, on Measures to Promote the Knowledge Society in Andalucia. 

[100] De Boor, A. Pmake. A tutorial, 

[101] De Icaza, M. "The story of the GNOME Project". 

[102] Senate of the Republic of France. Forum sur la proposition de loi tendant a 
generaliser dans l'administration l'usage d'Internet et de logiciels libres. 

[103] De las Heras Quiros, P.; Gonzalez Barahona, J. M. (2000). "Iniciativas de las 
administraciones publicas en relacion al software libre". Bole. TIC, ASTIC magazine (14). 

[104] Debian. "Debian free software guidelines". 


[105] Debian. Debian policy manual. 

[106] Debian. "Debian social contract". 

[107] Schriftenreihe der KBSt (2003, July). Leitfaden fur die migration von basissoft- 
warekomponenten auf serverund arbeitsplatzsystemen. Technical report, Koordinierungs- 
und Beratungsstelle der Bundesregierung fiir Informationstechnik in der Bundesverwaltung 

[108] DiBona, C; Ockman, S.; Stone, M. (ed.) (1999). Open sources. Voices from the open 
source revolution. O'Reilly & Associates. 

[109] Open Directory Project, 

[110] Ehrenkrantz, J. R. (2003, May). "Release management within open source projects". 
In: Proceedings of the 3 r Workshop on Open Source I 
Conference on Software Engineering. Portland, USA 

In: Proceedings of the 3' Workshop on Open Source Software Engineering at the 25' International 

[111] European Council (1991). Council Directive 91/250/CEE of 14 th May 1991, on the 
legal protection of computer programs. 

GNUFDL" PID_001 48386 195 Free Software 

[112] Feller, J.; Fitzgerald, B; Hissam, S.; Lakhani, K. (ed.) (2003). Making sense of the 
bazaar. O'Reilly. 

[113] Fogel, K.; Bar, M. (2001). Open source code development with CVS (2 nd edition). 
Paragliph Press. 

[114] Electronic Frontier Foundation. Open Audio. 

[115] Free Software Foundation. GPLv3. 

[116] Free Software Foundation. LGPLv3. First discussion draft. 

[117] Free Software Foundation (1985): "The GNU Manifesto". 

[118] Free Software Foundation (1991, junio). GNU General Public License, version 2. 

[119] Free Software Foundation (1999, February). GNU Lesser General Public License, 
version 2.1. 

[120] Free Software Foundation. "Free software definition". 

[121] Free Software Foundation. "Free licenses". 

[122] Garbee, B.; Koptein, H.; Lohner, N.; Lowe, W.; Mitchell, B.; Murdock, I.; 
Schulze, M.; Small, C. "A brief history of Debian". In the package: Debian-history. 

[123] German, D. (2002, May). "The evolution of GNOME". In: Proceedings of the 2 nd Work- 


shop on Open Source Software Engineering at the 24 International Conference on Software Engi- 
neering. Florida, USA 

[124] German, D.; Mockus, A. (2003, May): "Automating the measurement of open source 
projects". In: Proceedings of the 3 r Workshop on Open Source Si 
International Conference on Software Engineering. Portland, USA 

projects". In: Proceedings of the 3' Workshop on Open Source Software Engineering at the 25' 

[125] Ghosh, R. A. (1998, March). "Cooking pot markets: an economic model for the trade 
in free goods and services on the Internet. First Monday, 3(3). 

[126] Ghosh, R. A.; Glott, R.; Krieger, B.; Robles, G. (2002). Free/libre and open source 
software: Survey and study. Part iv: "Survey of developers". 

[127] Ghosh, R. A.; Prakash, V. V. (2000, July). "The orbiten free software survey". First 
Monday, 5(7). 

[128] Godfrey, M. W.; Tu, Q. (2000, August). "Evolution in open source software. A case 
study". In: Proceedings of the 2000 International Conference on Software Maintainance. 

GNUFDL» PID_001 48386 196 Free Software 

[129] Gonzalez, J. A. (2002, March). "Carta al congresista Villanueva". pe/mscarta.html 

[130] Goosens, M.; Rahtz, S. (1999). The LaTeX Web Companion. Addison Wesley. 

[131] Grad, B. (2002, January-March). "A personal recollection: IBM's unbundling of soft- 
ware and services". In: IEEE Annals of the History of Computing, 24(1):64-71. 

[132] Working Group on Libre Software (1999). "Free software / open source. Informa- 
tion society opportunities for Europe?". 

[133] GrULIC. "Legislation on the use of free software by the State ". 

[134] Hamerly, J; Paquin, T.; Walton, S. (1999). "Freeing the source. The story of Mozil- 

[135] Hammel, M. J. (1991, December). "The history of xfree86". Linux Magazine. 

[136] Harris, S. (2001, August). The Tao oflETR A novice's guide to the Internet engineering 
task force (RFC 3160, FYI 17). 

[137] Harrison, P. (2002). "The rational street performer protocol". 

[138] Hasan, R. "History of Linux". 

[139] Hauben, M.; Hauben, R. (1997). Netizens. On the history and impact of Usenet and the 
Internet. IEEE Computer Society Press. 

[140] Healy, K.; Schussman; A. (2003, January). "The ecology of open source software 

[141] Hecker, F. (1998, May). "Setting up shop. The business of open-source software". 

[142] Hecker, F. (1998). "Setting up shop. The business of open-source software". 

[143] Hertel, G.; Niedner, S.; Herrmann, S. (2003). "Motivation of software developers 
in open source projects. An Internet-based survey of contributors to the Linux kernel". 

[144] Himanen, P. (2001). The hacker ethic and the spirit of the information age. Random 

http : // w ww. 

[145] Hunt, F.; Johnson, P. (2002). "On the Pareto distribution of SourceForge projects. 
Technical report". Centre for Technology Management, Cambridge University Engineering 
Department, Mill Lane, Cambridge CB2 1RX. 

[146] Open Source Initiative. "History of the OSI". 

GNUFDL» PID_001 48386 197 Free Software 

[147] Hamilton, J. R. (US ambassador to Peru) (2002, June). "Carta al presidente del Con- 
greso de la Republica". 

[148] Jones, P. (2000, May). "Brook's law and open source. The more the merrier?". 

[149] Jorgensen, N. "Incremental and decentralized integration in FreeBSD". In: Feller etal. 

[150] Brooks, F. P. (1975). The mythical man-month. Essays on software engineering. Addison- 

[151] Kalt, C. (2000, April). "Internet relay chat: architecture (RFC 2810)". 

[152] Kelsey, J.; Schneier, B. (1998, November). "The street performer protocol". In: Third 
USENIX Workshop on Electronic Commerce Proceedings. USENIX Press. 

[153] Kelsey, J.; Schneier, B. (1999, June). "The street performer protocol and digital copy- 
rights". First Monday, 4(6). 

[154] Kelty, C. M. (2001, December). "Free software/free science". First Monday, 6(12). 

[155] Khatib, J. "OpenlPCore Hardware General Public License". 

[156] Knuth, D. (1989). The TeXbook. Addison Welsley. 

[157] Koch, S. (ed.) (2003). Free/open source software development. Idea Group Inc. 

[158] Koch, S.; Schneider, G. (2000). "Results from software engineering research into 
open source development projects using public data". In: Diskussionspapiere zum Tdtigkeitsfeld 
Informationsverarbeitung und Informationswirtschaft, H.R. Hansen und W.H. Janko (Hrsg.), Nr. 
22, WirtschaftsuniversitatWien. 

[159] Kovacs, G. L.; Drozdik, S.; Succi, G.; Zuliani, P. (2004). "Open source software 
for the public administration". In: Proceedings of the 6 l International Workshop on Computer 
Science and Information Technologies (CIST 2004). Budapest, Hungary. 

[160] Krishnamurthy, S. (2002, May). "Cave or community? An empirical examination of 
100 mature open source projects". First Monday, 7(6). 

[161] Laffitte; Tregouet; Cabanel (1999). Proposition de loi numero 495. Senate of the 
Republic of France. 

[162] Laffitte; Tregouet; Cabanel (2000). Proposition de loi numero 117. Senate of the 
Republic of France. 

[163] Lamport, L. (19 

Welsley, Reading, Mass. 

[163] Lamport, L. (1994). LaTeX user's guide and reference manual (2 nd edition). Addison 

GNUFDL» PID_001 48386 198 Free Software 

[164] Lancashire, D. (2001, December). "Code, culture and cash. The fading altruism of 
open source development". First Monday, 6(12). 

[165] Lehman, M. M.; Ramil, J. F; Wernick, P. D. (1997, November). "Metrics and laws 

of software evolution. The nineties view". In: Proceedings of the 4 U International Symposium 
on Software Metrics. 

[166] Leiner, B. M.; Cerf, V. G.; Kahn, R. E.; Clark, D. D.; Kleinrock, L.; Lynch, 

D. C; Postel, J.; Roberts, L. G.; Wolff, S. (1997). "A brief history of the Internet". In: 
Communications of the ACM. 

[167] Netcraft Ltd. August 2003 Web Server Survey, 2003. 

[168] Lucovsky, M. (2000). "From NT OS/2 to Windows 2000 and beyond. A software-en- 
gineering odyssey". _html/> 

[169] McGraw, G. "Building secure software: how to avoid security problems the right way". 
Cited by: David A. Wheeler in 

[170] McKusick, M. K. (1999). "Twenty years of Berkeley Unix. From AT&T owned to freely 
redistributable". In: DiBona era/. [108]. 

[171] SUN Microsystems (2000). "Sun microsystems announces availability of StarOffice 
source code on". 

[172] Mockus, A.; Fielding, R. T.; Herbsleb, J. D. (2000, June). "A case study of open 
source software development: the Apache server". In: Proceedings of the 22 m International Con- 
ference on Software Engineering (ICSE 2000), pages 263272. Limerick, Ireland ACM Press. 

[173] Molenaar, B. "What is the context of charityware?". 

[174] MIT OpenCourseWare. 

[175] Nagel, L. W. (1996, September). "The life of SPICE". In: 1996 Bipolar Circuits and Tech- 
nology Meeting. Minneapolis, MN, US 

[176] Narduzzo, A.; Rossi, A. (2003, May). "Modularity in action: GNU/Linux and free/ 
open source software development model unleashed". 

[177] Newman, N. (1999). "The origins and future of open source software". 

[178] Nupedia. 

[179] Villanueva Nunez, E. (2002, April). "Letter to Microsoft Peru". 

GNUFDL" PID_001 48386 199 Free Software 

[180] Danish Board of Technology (2002, October). "Open-source software in e-Govern- 
ment, analysis and recommendations drawn up by a working group under the danish board 
of technology. Technical report". 

[181] Open Source Initiative. "Open source licenses". 

[182] Pareto, W. (1896). "Course of Political Economy". Lausanne. 

[183] Perens, P.; The Open Source Initiative (1998). "The open source definition", http:/ 

[184] GNU Peru. "Proyectos ley de software libre en la Administration piiblica del Gobierno 
peruano, Congreso de la Republica". 

[185] Pinheiro, P. (1999, December). Proposicao pl-2269/1999: Dispoe sobre a utilizacao de 
programas abertos pelos entes de direito publico e de direito privado sob controle acionario 
da administracao piiblica. Camara dos Deputados do Brasil. 

http://www.camara. asp?id=l 7879 

[186] Pranevich, J. (2003). "The wonderful world of Linux 2.6". 

[187] The Debian Project. "Debian developer map". 

[188] Puigcercos Boixassa, J. (2002). Draft Bill on Measures for Implementing Free Soft- 
ware in Public Administration. 


[189] Quittner, J.; Slatalla, M. (1998). Speeding the net: the inside story of Netscape and how 
it challenged Microsoft. Atlantic Monthly Pr. 

[190] Rasch, C. "A brief history of free/open source software movement". 

[191] Rasch, C. (2001, May). "The Wall Street performer protocol. Using software completion 
bonds to fund open source software development". First Monday, 6(6). 

[192] Raymond, E. R. (2001, January). The cathedral and the bazaar. Musings on Linux and 
open source by an accidental revolutionary. O'Reilly & Associates ( 

[193] Reis, C R.; De Mattos Fortes, R. P. (2002, February). "An overview of the software 
engineering process and tools in the Mozilla Project". 

[194] Rideau, F. R. (2000). "Patents are an economic absurdity". 

[195] Roberts, L. (1978, November). "The evolution of packet switching". Proceedings of the 
IEEE, (66). 

[196] Robles, G.; Gonzalez Barahona, J. M.; Centeno Gonzalez, J.; Matellan Oliv- 
ers, V.; Rodero Merino, L. (2003, May). "Studying the evolution of libre software projects 

GNUFDL» PID_001 48386 200 Free Software 

using publicly available data". In: Proceedings of the 3 r Workshop on Open Source Software En- 
gineering at the 25' International Conference on Software Engineering. Portland, US. 

[197] Robles, G.; Scheider, H.; Tretkowski, I.; Weber, N. (2001): "Who is doing it? 
Knowing more about libre software developers". 

[198] Rochkind, M. (1986, May). "Interview with Dick Haight". Unix Review. 

[199] Scacchi, W. (2003). "Understanding open source software evolution. Applying, break- 
ing and rethinking the laws of software evolution". 

[200] Schneier, B. (2000). "Software complexity and security". 

[201] Smoogen, S. J. "The truth behind Red Hat names". 

[202] Haggen So. "Comparison of free/open source hosting (FOSPhost) sites available for 
hosting projects externally from project owners". 

[203] Stallman, R. "GNU coding standards". 

[204] Stall man, R. "Why free software is better than open source". 

[205] Stallman, R. (1998). "Copyleft: pragmatic idealism". 

[206] Stallman, R. (1998). "Why free software is better than open source". 

[207] Stallman, R. (1998). "Why software should not have owners". 

[208] Stallman, R. "The GNU Operating System and the Free Software Movement". In: 
DiBonaeta/. [108], 

[209] Stallman, R. (1999, June). "On free hardware". Linux Today. 

[210] Stallman, R. (2001). "The free universal encyclopedia and learning resource". 

[211] Stallman, R. (2002). Free software, free society. Selected essays of Richard M. Stallman. 
Joshua Gay. 

[212] Stallman, R. (2003). "Some confusing or loaded words and phrases that are worth 

[213] Stoltz, M. (1999). "The case for government promotion of open source software". 

GNUFDL" PID_001 48386 201 Free Software 

[214] Tanenbaum, A.; Torvalds, L. (1999). "The Tanenbaum-Torvalds debate". 

[215] The Open Source Initiative. "The open source definition". 

[216] Tiemann, M. "Future of Cygnus Solutions. An entrepreneur's account". In: DiBona 
etal. [108]. 

[217] Torvalds, L; Diamond; D. (2001). Just for fun: the story of an accidental revolutionary. 

[218] Linus Torvalds, Hamano, J. C; Ericsson, A. "Git manual page". 

[219] Tuomi, I. (2002). "Evolution of the Linux credits file: methodological challenges and 
reference data for open source research". 

[220] Several authors. "Open letter to WIPO". 

[221] Vigo i Salient, P.; Benach i Pascual, E.; Huguet i Biosca; J. (2002, May). Proposi- 
cio de llei de programari lliure en el marc de l'Administracio publica de Catalunya. 

modules. php?op=modload&name=Sections&file=index&req=viewarticle&artid=49 

[222] Villanueva Nunez, E. (2001, December). Free software project bill, number 1609. 

[223] Villanueva Nunez, E.; Rodrich Acker man, J. (2002, April). Bill on the use of free 
software by the Public Administration, number 2485. 

[224] W3C (2000). Extensible markup language (xml) 1.0 (2 nd edition). 

[225] Walsh, N.; Muellner, L.; Stayton, B. (2002). DocBook: the definitive guide. O'Reilly. 

[226] Welke, L; Johnson, L. (1998). How the ICP Directory began. 

[227] Wheeler, D. A. (2000, July). "Estimating Linux's size". 

[228] Wheeler, D. A. (2001, June). "More than a gigabuck: estimating GNU/Linux's". 

[229] Wiesstein, E. "Concise encyclopedia of mathematics". 

[230] Wikipedia. "Gini coefficient". 

GNUFDL" PID_001 48386 202 Free Software 

[231] Wikipedia. "Lorenz curve". 

[232] Wikipedia. "Pareto". 

[233] Wikipedia. "TeX". 

[234] Wilson, B. "Netscape Navigator". 

[235] Computer World (2000). "Salary survey 2000". 

[236] Young, R. (1999). "Giving it away, how Red Hat software stumbled across a new eco- 
nomic model and helped improve an industry". 

[237] Zawinsky, J. W. (1999). "Resignation and postmortem". 


Jesus M. Gonzalez Barahona 
Joaquin Seoane Pascual 
Gregorio Robles 



Universitat Oberta 
de Catalunya 

GNUFDL • PID_001 48385 Appendixes 

Copyright © 2010, FUOC. Permission is granted to copy, distribute and modify this document either under the terms of the GNU 
Free Documentation Licence, Version 1 .2 or any subsequent version published by the Free Software Foundation, with no invariant 
sections or front-cover or back-cover texts, or under the terms of Creative Commons by-sa 3.0 license, at the option of the user. A 
copy of these licenses is included in the corresponding appendixes of this document. 

GNUFDL" PID_001 48385 Appendixes 


1. Appendix A. Learning guide 5 

2. Appendix B. Key dates in the history of free software 10 

3. Appendix C. GNU Public License 17 

4. Appendix D. Texts of some legislative proposals and 

related documents 25 

5. Appendix E. Creative Commons' Attribution-Share Alike 56 

6. Appendix F. GNU Free Documentation License 64 

GNUFDL» PID_001 48385 5 Appendixes 

1. Appendix A. Learning guide 

A.l. Introduction 

What is free software? What is it and what are the implications of a free pro- 
gram licence? How is free software developed? How are free software projects 
financed and what are the business models associated to them that we are 
experiencing? What motivates developers, especially volunteers, to become 
involved in free software projects? What are these developers like? How are 
their projects coordinated, and what is the software that they produce like? 
In short, what is the overall panorama of free software? 

These are the sort of questions that we will try to answer in this document. 
Because although free software is increasing its presence in the media and 
in debates among IT professionals, and although even citizens in general are 
starting to talk about it, it is still unknown for many people. And even those 
who are familiar with it are often aware of just some of its features, and mostly 
ignorant about others. 

A.2. Aims 

The general aim is, unquestionably, that the reader understand and think log- 
ically about basic free software concepts and their main implications. Let us 
look for more specific aims: 

• Knowing what is (and what is not) free software and the main conse- 
quences that such a definition has. 

• Exploring the rudiments of the legal questions surrounding free software 
and, particularly, the importance of licenses, the main types and their con- 

• Having a perspective of the reality of free software, from a global and his- 
torical point of view and from the perspective of the most advanced and 
current projects. 

• Learning and getting to know the methods in which free software projects 
may be financed (when such means exist) and the relevant business mod- 

• Learning the most important details of the free software development 
models and the methods for studying them from the perspective of soft- 
ware engineering. 

GNUFDL* PID 00148385 


A. 3. Contents and learning plans 

This text is structured into various chapters (didactic modules) and written in 
such a way that they are practically independent and self-contained, which 
means that, excepting the introduction, the book can be read in any order. 
However, readers are advised to follow the order established for the book, in 
accordance with the plan below. 

The course will be structured in ECTS credits, which means that the planning 
will require an overall effort form the student, which will include exercises 
and debates, which will last 150 hours. 

Chapter 1 (6 hours). Introductory module discussing all the specific aspects 
of free software and focusing essentially on an explanation of the underlying 
basis, for people who are learning about the matter for the first time and on 
highlighting its importance. An introduction covering the definition of free 
software and its main consequences, amongst other elements, will be provid- 






Learning what freedom means with re- 
gard to software 

The four freedoms 


Reading the material 


Distinguishing between free software 
and other related concepts 

Definition of related concepts, whether 
they are similar or analogous 


Reading the material and making 

1 hour 

Introducing the reasons for which free 
software is made 

Ethical and practical motivations 


Reading the material and making 

1 hour 

Introducing the consequences of free 

Consequences for the user, the State, 
the developer, etc. 


Reading the material and making 


Chapter 2 (14 hours). Historical development of the world of free software, 
from its beginning in the seventies to the current moment, offering a broad 
vision of the most notable milestones, the main projects, the financial, pro- 
fessional or social evolution, etc. 






Learning about the "prehistory" 
of free software 

Facts before the existence of 
the concept 

Section 2.1 and beginning 
of annex B 

Reading the material and 
making suggestions 


Learning about the history of free 
software all the way up to the 
present day 

Most significant events in 
chronological order 

Sections 2.2, 2.3, 2.4 and 
rest of annex B 

Reading the material and 
making suggestions 


Trying to predict the future 

Some predictions (hopes and 

Section 2.5 

Reading the material and 
making suggestions 


GNUFDL* PID 00148385 


Chapter 3 (9 hours). Legal aspects of free software. The most common free 
software licenses and their effects on business and development models will 
be analysed in detail. 






Learning the basic concepts of intellec- 
tual and industrial property 

Copyright, intellectual property, 
patents, brands, industrial secrets 


Reading the material and making 


Learning the legal basis of free software: 
the licenses 

Definition of free licenses and the fea- 
tures of the most important licenses 


Reading the material and making 


Chapter 4 (8 hours). Characteristics of free software developers and the moti- 
vations that lead them to participate in the projects, thereby making the ex- 
istence of free programs possible. 






Getting to know the type of people 
that develop free software 

Ages, genders, professions, geo- 
graphical location, etc. 

Sections 4.1, 4.2, 
4.3 and 4.4 

Reading the material and 
making suggestions 


Learning how much time to spend 
on it and why 

Weekly dedication, motivations, 
questions of prestige and leader- 

Sections 4.5, 4.6, 
4.7 and 4.8 

Reading the material and 
making suggestions 


Chapter 5 (22 hours). Financial aspects of free software and, especially meth- 
ods for financing the projects and business models that are being explored. 






Learning about the sources of fi- 

Financial sources used 

Section 5.1 

Reading the material and mak- 
ing suggestions 


Learning how to profit from free 

Business models 

Sections 5.2 and 5.3 

Reading the material and mak- 
ing suggestions 


Learning about the relationship be- 
tween free software and the monop- 
olistic situations that are typical in 
the software industry 

Monopolies and software. Free 
software's role 

Sections 5.1, 5.2, 
5.3 and 5.4 

Reading the material and mak- 
ing suggestions 


Chapter 6 (28 hours). Relationship of policies and free software and, especially 
policies for promoting free software and the use of free software by public 






Learning about the effect of free soft- 
ware on public administrations. 

Main effects and difficulties in imple- 


Reading the material and making 


Learning about what administrations do 
or can do with regard to free software 

Solutions to needs, promotion and in- 
vestment in R&D 


Reading the material and making 


GNUFDL* PID 00148385 







Learning about legislative initiatives 

Revision of legislative initiatives for im- 
plementing or supporting free software, 
including examples of specific texts. 


Reading the material and making 


Chapter 7 (12 hours). Management and development models for free software 
projects, techniques that have been successful and quantitative and qualita- 
tive studies of free software from the perspective of development. 






Learning about the paradigmatic 
models of software development 

"The cathedral and the 

Sections 7.1, 7.2, 7.3 
and 7.5 

Reading the material and suggested 


Learning about the processes involved 
in the development of free software 

Characteristic processes 

Section 7.4 

Reading the material and suggested 


Learning about the possibilities and re- 
alities that the availability of sources 
and the associated registries bring to 
the free software engineering 

Resources and quantita- 
tive studies 

Section 7.6 

Reading the material and suggested 


Learning what remains to be done in 
free software engineering 

Future tasks 

Section 7.7 

Reading the material and suggested 


Chapter 8 (14 hours). Introduction of the technologies and development en- 
vironments for free software and their effects on the management and evolu- 
tion of the projects. 






Learning the general features of the 
environments and the tools that free 
software developers use 

General characterisation 

Section 8.1 

Reading the material and suggest- 
ed bibliography 


Learning the basic development tools 

Languages, compilers, operating 
systems, etc. 

Section 8.2 
and 8.3 

Reading the material and suggest- 
ed bibliography 


Learning the basic methods with 
which developers work together 

Messaging, forums, repositories, 
chats and wikis 

Section 8.4 

Reading the material and suggest- 
ed bibliography 


Learning the mechanisms used to 
manage sources and their versions 

CVS and new alternatives 

Section 8.5 

Reading the material and suggest- 
ed bibliography 


Learning how free software is docu- 

Languages and tools for documen- 

Section 8.6 

Reading the material and suggest- 
ed bibliography 


Learning how errors and tasks are 

Bug management systems 

Section 8.7 

Reading the material and suggest- 
ed bibliography 

1 hour 

Learning how portability is supported 

Resources for other architectures 

Section 8.8 

Reading the material and suggest- 
ed bibliography 


Learning about the public environ- 
ments of integrated development 

SourceForge and others 

Section 8.9 

Reading the material and suggest- 
ed bibliography 


GNUFDL* PID 00148385 


Chapter 9 (30 hours). Studying free software projects (revising the most in- 
teresting classical free software projects, in terms of results obtained, manage- 
ment model, historical evolution, effect on other projects, etc.). Study of com- 
panies related to free software. 






Learning an example of operating sys- 

Linux and *BSD 

Sections 9.1 and 

Reading the material and suggested bibliog- 


Learning an example of desktop envi- 

Gnome and KDE 

Sections 9.3 and 

Reading the material and suggested bibliog- 


Learning an example of system pro- 


Section 9.5 

Reading the material and suggested bibliog- 


Learning an example of end user pro- 

Mozilla and 

Sections 9.6 and 

Reading the material and suggested bibliog- 


Learning an example of a distribution 

Red Hat and De- 

Sections 9.8 and 

Reading the material and suggested bibliog- 


Chapter 10 (6 hours). Module in which free resources other than software 
are presented; these are resources that have been created partly thanks to free 
software and the model that it has given. 






Learning other free resources 

Free texts, hardware, teaching materials 
and art 


Reading the material and suggested 


Learning about the applicable 

Licenses, especially the Creative Com- 
mons licenses 


Reading the material and suggested 


GNUFDL' PID 00148385 


2. Appendix B. Key dates in the history of free 

This is only a list of the dates that could be considered to be important in the 
history of free software. It is based on the one that appears in [132] and the 
one provided by the Open Source Initiative [146] and is not supposed to be 
comprehensive: there are certainly many important dates that have not been 
included in the list. However, we hope to provide a sufficiently complete view 
of the historical landscape in which the world of free software has evolved. 




1950s and 1960s 

The software is distributed with its source code and without any restrictions on the user groups 
such as SHARE (IBM) and DECUS (DEC). 

1 969, April 

RFC number 1, which describes the first Internet (then called ARPANET) is published. The free 
availability of the RFCs and, particularly, of the specifications of the protocols used in Internet 
were key factors for its development. 

1970, January 

IBM began selling its software separately, creating the beginning of the proprietary software in- 


Unix begins to be distributed in universities and research centres. 


Unix arrives at Berkeley University, in California. The history of Unix BSD begins. 


SPICE is placed by Donald O. Penderson in the public domain. With time, it will become the 
standard in its field (integrated circuit simulators). 


Donald Knuth, of Stanford University, starts working on TeX, an electronic typesetting system 
that will be distributed as free software. 


Richard Stallman writes "The GNU Manifesto", in which he asks for software to be shared with 
the public again. 


The GNU project begins. The developers that work on it, initially coordinated by Richard Stall- 
man, begin to create a large number of tools similar to those in Unix, including an editor 
(Emacs) and a compiler (GCC). The aim is to build an operating system that is completely free. 


The X Consortium, based at MIT, distributes the X Window system as free software, under a li- 
cense that is hardly restrictive at all. 


Richard Stallman founds the Free Software Foundation. Among other tasks, the Foundation will 
work as a centre that receives the funds and resources that will assist the development of the 
GNU project and as the owner of the intellectual property generated by the project. 


Cygnus, the first company that essentially provides commercial services for free software (in- 
cluding support, development and adaptation of free programs), is founded. 


The Network Simulator (or simply, ns) begins to be developed as a variant of the REAL Network 
Simulator. Ns is a free telecommunication network simulator that will be used extensively by 
universities all over the world and that will become a standard in its field, to a certain extent. 


The Free Software Foundation announces that it intends to build a kernel that will be called 
GNU Hurd. The aim of this project is to complete what the GNU project's strategy was most 
missing: a complete operating system. 

GNUFDL' PID 00148385 






William and Lynne Jolitz write a series in Dr. Dobbs journal on how to port BSD Unix to PC based 
on the i386. 

1991, August 

Linus Torvalds, a twenty-one year old Finnish student announces that he has begun work on a 
free Unix-type kernel using GNU tools, such as GCC. His aim at the time is to build a free Minix. 

1991, October 

Linus Torvalds releases the first version of his kernel, which is still very primitive and is called Lin- 


The US Air Force awards New York University a contract to build an open source compiler for 
the new version of Ada (a language that it was almost obligatory to use at that time in all con- 
tracts with the US military), Ada 95. The NYU team chooses GNU GCC for the generation of 
code and calls its compiler GNAT (GNU NYU Ada 95 Translator). 

1992, July 

William and Lynne Jolitz release 386BSD 0.1, which, with time, will give rise to the projects 
NetBSD, FreeBSD and later OpenBSD. 


SuSE is founded in Germany, which begins its business distributing Slackware Linux, translated 
into German. 

1993, August 

Ian Murdock starts a new distribution based on Linux called Debian GNU/Linux, which will be- 
come the distribution built by voluntary developers with the most participants. 

1993, December 

FreeBSD 1 .0, one of the first stable distributions derived from the Jolitz's 386BSD is released on 
the Internet. 


The GNAT developers found the company Ada Core Technologies, with the aim of guarantee- 
ing its development and evolution in the future and with a business model based on providing 
services to their clients' compiler (and not selling the compiler itself, which continues to be free 
software). With time, GNAT will become the leader in the market of Ada compilers. 

1994, January 

Version 0.91 of Debian GNU/Linux is released; it is the fruit of the efforts of twelve developers. 

1 994, March 

The first edition of the Linux journal is published. 

1994, 29 th July 

Marc Ewing publishes the first version of Red Hat Linux. As is the case with Debian, the aim is to 
improve the results of the predominant distribution in that time, Slackware. 

1 994, October 

NetBSD 1 .0. is released 


Bob Young founds Red Hat Software buying the Red Hat Linux distribution from its creator, 
Marc Ewing, and merging it with his own business, ACC, which has been selling materials relat- 
ed to Linux and Unix through catalogue since 1993. A little later, Red Hat Linux 2.0 is released; 
it is the first distribution that includes the RPM packaging format. 


DARPA supports the development of ns through the VINT project. 

1995, January 

FreeBSD 2.0. is released 

1 995, April 

The first official release of Apache (0.6.2) takes place. 


The First Conference on Freely Redistributable Software takes place in Cambridge, Mas- 
sachusetts, US. 

1 996, October 

The KDE project is announced; it is one of the first to address usability problems in the Unix en- 
vironment and the first that tries to do so on a large scale in the world of free software. 

1997, January 

Eric S. Raymond presents his paper "The cathedral and the bazaar", in which he expresses his 
opinions on why certain free software development models work. 

1997, August 

Miguel de Icaza announces the GNOME project, a competitor to KDE with similar aims, but with 
the explicit objective of ensuring that the whole of the resulting system is free software. Born as 
a reaction of the Free Software Foundation and others to the licensing problems that KDE had, 
which involved a fundamental component, the Qt library, which was not free software at that 

GNUFDL' PID 00148385 





1 998, 22 nd January 

Netscape declares its intention of distributing as free software the code of its browser (Netscape 
Navigator), which had been the leader in the web browser market. 

1 998, 3 rd February 

Chris Peterson, Todd Anderson, John Hall, Larry Augustin, Sam Ockman and Eric Raymond meet 
up to study the consequences of Netscape's announcement with regard to the release of its 
browser and decide to promote the term open source software [146], using it as a brand that 
guarantees that the products that have it consist of free software. The promoters of this term 
understand that it is more appropriate for the corporate world than the one that was more 
commonly used up to that moment, free software. The Open Source Initiative is created to man- 
age the term. 

1998, 31 st March 

Netscape publishes a large part of its source code for Netscape Navigator on the Internet. 

1 998, 7 th May 

Corel announces the NetWinder, a network computer based on Linux. It is the first time that 
a large company commercialises an element that uses software that is basically free software. 
Shortly afterwards, Corel announces its plan to port its office software (which includes WordPer- 
fect) to Linux, which is also a novelty for the time. 

1998, 28 th May 

Sun Microsystems and Adaptec become part of Linux International. They are the first big IT 
companies to do so. 

1998, June 

The technical conference of USENIX, which is usually dedicated to Unix, opens a parallel session 
called FREENIX, focusing on free software. 

1998, 22 nd June 

IBM announces that it will commercialise and provide support for Apache, using it as the server 
of its WebSphere product line. 

1998, July 

Debian GNU/Linux 2.0 is released; it has been built by more than three hundred volunteers and 
the distribution includes more than one thousand five hundred packages. 

1998, July 

KDE 1 .0 is released; it is the first version distributed as stable. Several GNU/Linux distributions in- 
corporate it shortly afterwards. 

1998, August 

Linus Torvalds and Linux appear on the cover of Forbes magazine. 

1998, 29 th September 

Red Hat, which is the leading company in the market of Linux-based distributions at the time, 
announces that Intel and Netscape have bought a minority share in its capital. Free software be- 
gins to awaken interest among investors. 

1 998, November 

MandrakeSoft is founded and shortly afterwards, it releases Mandrake Linux, its distribution of 

1998, 1 st November 

The Halloween Documents, in which Microsoft supposedly identifies GNU/Linux and free soft- 
ware as an important competitor and plans how to attack it, are published. 

1999, 27 th January 

HP and SGI announce that they will support Linux in their computers, which marks the begin- 
ning of a trend: the abandonment of proprietary Unix by the computer manufacturers that used 
them as their operating system, in favour of Linux. 

1 999, March 

GNOME 1 .0, which will subsequently be made more stable (October GNOME) and incorporat- 
ed in several GNU/Linux distributions, is released. 

1 999, 9 th March 

Debian GNU/Linux 2.1 is released, with more than two thousand packages. 

1 999, 1 5 th March 

Apple releases Darwin, which will be the central component of its new Mac OS X, under a free 

1 999, August 

Red Hat is floated on the stock exchange. The price of the shares increases enormously in the 
first days after the float, to the extent that it is capitalised at 4,800 million dollars. Later, other 
companies related to free software, such as VA Linux and, will also be floated on 
the stock exchange. The value of the shares of all these companies will plummet a few years lat- 
er, when the dotcom bubble explodes; many of these companies will not survive the event. 

GNUFDL' PID 00148385 





1 999, October 

Two companies are founded in order to produce software in the framework of the GNOME 
project: Eazel (which will go bankrupt in 2002, after producing Nautilus, a file manager) and 
Helix Code (later renamed Ximian and subsequently bought by Novell, which will produce tools 
such as Red Carpet or Evolution). 

1 999, November 

Red Hat Software buys Cygnus. The resulting company is the biggest company in the world in 
the field of free software. 

2000, January 

Mozilla M1 3, considered by many as the first reasonably stable version of Mozilla, is released al- 
most two years after the release of a large part of Netscape Navigator's code. 

2000, May 

GNOME 1 .2 (Bongo GNOME) is released. 

2000, August 

The creation of the GNOME Foundation is announced. 

2000, 1 5 th August 

Debian GNU/Linux 2.2 is released, with more than two thousand five hundred source packages, 
which comprise approximately 55 million lines of code. 

2001, January 

Version 2.4 of Linux is released. 

2001, 15 th January 

Wikipedia is started. The idea of building an encyclopaedia using a wiki as IT support, where, in 
principle, anyone can cooperate, applying working methods that are very similar to those used 
in free software, becomes a reality. 

2002, 30 th January 

ObjectWeb, an organisation founded in France by Bull, France Telecom and INRIA that is one of 
the first organisations designed to produce free software by cooperating with companies and re- 
search centres, is founded with clearly commercial objectives and the idea of being the nucleus 
of an international community of interests. 

2002, 3 rd April 

KDE 3.0, the third generation of the KDE desktop environment, is released. The quality of free 
desktops begins to match that of traditional commercial desktops. 

2002, April 

The gnuLinEx project is publicly announced; with this project, the Regional Government of Ex- 
tremadura (Spain) wishes to use its own GNU/Linux distribution in the computers of all the pub- 
lic schools in the region. 

2002, May 

Mozilla 1 .0, the first officially stable version of the project, is released. 

2002, 1 st May 

The office suite, 1 .0, is released; it will soon become a standard office applica- 
tion suite in the free software world. 

2002, 26 th June 

GNOME 2.0, which represents an important step forward for users, with a more carefully de- 
signed interface and more attention to user-friendliness, is released. Other aspects that improve 
the accessibility are also introduced. 

2002, 19 th July 

Debian GNU/Linux 3.0 is released with more than 100 million lines of source code; more than 
nine hundred developers participate in this version. 

2002, 28 th July 

Version 3.0 of Knoppix is released; it is an evaluation distribution that can be installed on a hard 
disk quickly and easily, and it becomes a tremendous success. 

2002, 23 rd September 

The first version of Firefox (which is called Phoenix at the time) is released, as an experimental 
extension based on the code of Mozilla Suite that is supposed to be simpler. 

2002, December 

Red Hat Softwre announces that its cash flow in the second and third quarters of 2002 was pos- 

2002, 1 6 th December 

The first Creative Commons licenses are published (although the project was launched in 2001 ). 

2003, January 

MandrakeSoft, a company that produces the Mandrake Linux distribution, declares bankruptcy. 

2003, 19 th January 

FreeBSD 5.0-RELEASE is released, after almost three years of work since the previous stable large- 
scale version. 

GNUFDL' PID 00148385 





2003, 22 nd January 

The number of articles in English on Wikipedia reaches one hundred thousand articles. Shortly 
afterwards, the number of German articles reaches ten thousand. 

2003, February 

Motorola begins selling the A760 in China; it is the first mobile telephone that uses an operating 
system based on Linux (the MontaVista Linux distribution). 

2003, 6 th March 

The SCO group files a lawsuit against IBM for devaluing its version of Unix. This marks the begin- 
ning of a lawsuit in which IBM is accused of contributing code that belongs to SCO to the Linux 

2003, 28 th May 

Munich City Council (Germany) announces that Linux will replace Windows in most of its com- 
puter systems. 

2003, July 

MandrakeSoft announces that its finances have been positive for the whole year and that it ex- 
pects to come out of receivership in late 2003. 

2003, 7 th July 

An open letter [220] is written to the WIPO (World Intellectual Property Organization) asking it 
to examine new open models of collaborative creation (including free software but also the Hu- 
man Genome project or open scientific journals). 

2003, 15 th July 

The Mozilla Foundation is established. Netscape Inc. (now the property of AOL) announces that 
it will no longer develop the Netscape browser and, therefore, it will no longer work on the 
Mozilla project. The Mozilla Foundation is established with a donation of two million dollars 
from AOL and material support and human resources from various companies, including AOL it- 
self, Red Hat and Sun Microsystems. 

2003, 4 th August 

Novell buys Ximian Inc., one of the leading companies in the development of free software (es- 
pecially for GNOME), as part of its strategy to establish itself in the market for Linux-related so- 

2003, 2 nd September 1.1 is released. 

2003, 24 th September 

The European Parliament amends the Directive on Patentability of Computer-Implemented In- 
ventions so that (if it is approved as it stands) software patents are not allowed in the European 
Union. The Directive, which was originally proposed by the European Commission precisely to 
ensure that these types of patents were legal, is still in the codecision procedure, in which the 
Council of Ministers will also have to provide its opinion. 

2003, 5 th November 

Version 1 (FC1 ) of de Fedora Core, the fruit of the communal development process that Red 
Hat had announced a few months before, is released. As of this moment, the company Red Hat 
will commercialise Red Hat Enterprise Linux, whilst the Fedora Core collections are not officially 
maintained by Red Hat, but by the community of voluntary developers that build it with the as- 
sistance of Red Hat (which already existed before Red Hat decides on this collaboration). 

2004, 1 3 th January 

Novell finishes its purchase of SuSE for a total of 21 million dollars. 

2004, 9 th February 

The Mozilla Foundation decides to change the Mozilla Firebird name (previously called Phoenix) 
to Mozilla Firefox. This will be the definitive name of the browser, while its development is close 
to version 1 .0. 

2004, 1 8 th May 

The European Council, as part of the codecision process on the European Directive on the 
Patentability of Computer-Implemented Inventions, decides to submit a compromise version of 
the text to the European Parliament; however, it is accused of ignoring the Parliament's vote, as 
the new version permits the patenting of software. The decision is so contentious, even within 
the Council itself, that it is not formally approved until March 2005. 

2004, 8 th September 

Pepper Computer announces that it will launch the first miniPC with a touch screen that uses an 
operating system that is completely free, based on Fedora Core. 

2004, 20 th September 

The number of articles on Wikipedia reaches one million, in one hundred and five languages. 

2004, 20 th October 

The first version of Ubuntu is released; it is based on Debian and the aim is to publish new ver- 
sions regularly. The construction of the distribution is financed by the company Canonical, 
which offers maintenance and services for the distribution. The distribution will become very 
successful, fairly quickly. 

GNUFDL' PID 00148385 





2004, 9 th November 

Version 1 .0 of Firefox is released, after a long series of preparatory versions. This version was 
downloaded more than 25 million times in the one hundred days following its release. 

2005, 24 th January 

MandrakeSoft announces that it is buying the Brazilian company Conectiva, which releases a 
distribution based on linux with the same name. Shortly afterwards, MandrakeSoft announces 
that it is changing its name to Mandriva. 

2005, 1 st May 

OASIS recognises ODF (open document format), the data format use by 2.0, 
among others, as a standard. 

2005, 25 th May 

Nokia announces its Nokia 770, a miniPC that uses a version of Debian GNU/Linux with the X 
Window system and GTK+. 

2005, 6 th June 

Debian GNU/Linux 3.1 is released; it now has more than 200 million lines of source code. 

2005, 14 th June 

Sun Microsystems releases Open Solaris, the free version of its Solaris operating system. 

2005, 15 th June 

Mandriva buys the US company Lycoris (previously called Redmond Linux) and begins working 
on a distribution that incorporates the previous versions of Mandrake, Conectiva and Lycoris. 

2005, 6 th ly 

The European Parliament rejects the proposal of the Directive on the Patentability of Computer- 
Implemented Inventions received from the Council of Ministers, during the second reading. This 
means that the only legal text applicable to the subject in the European Union is the European 
Patent Convention of 1 973. 

2005, 20 th October 

Version 2.0 of, which is distributed under the LGPL, is released. 

2005, December 

The first version of Ruby on Rails, a work environment for the development of web applications 
using the model-view-controller architecture, is released. Distributed with license X1 1, it will be 
widely used in the prototyping and development of numerous web services. 

2005, December 

Nicholas Negroponte announces the OLPC (One Laptop Per Child) project, which has the aim 
of designing and building a portable PC of 1 00 dollars for children in developing countries. It 
uses free software with a GNU/Linux version called Sugar, based on Red Hat. 

2005, 14 th December 

The science journal Nature publishes a paper comparing Wikipedia with the Encyclopaedia Bri- 
tannica; according to the paper, the degree of precision with regard to scientific subjects of 
both encyclopaedias is similar. 

2006, 16 th January 

The first draft of the GPLv3 is published; it is an attempt to update the GPL, which is the license 
that is most commonly used for free software projects at the time (and by a long way). At this 
point, an open debating process begins with regard to the changes. 

2006, 1 st March 

The number of articles in English on Wikipedia reaches one million. 

2006, 20 th March 

Fedora Core 5 is released. 

2006, 1 st June 

Ubuntu 6.06 LTS is released; it is advertised as being supported by the company Canonical for 
three years. 

2006, August 

The number of Firefox downloads reaches 200 million (there are many more downloads from 
unofficial sites, which are not taken into account). Around this time, it is estimated that the 
browser has a 12% share of the global market (approximately 20% in Europe). 

2006, 12 th November 

Sun announces that it will release the different versions of the Java platform under the GPL. Up 
until this moment, these versions had been distributed for free in binary, which Sun had justified 
citing compatibility and stability issues; however this has made it extremely difficult to use Java 
in free software distributions. 

2006, 30 th November 

The ISO (International Standards Organization) and the IEC (International Electrotechnical Com- 
mission) jointly publish OASIS' ODF version as an international standard (ISO/IEC 26300:2006) 
for the exchange of office information. 

GNUFDL' PID 00148385 





2006, December 

The Taiwanese company First International Computer (FIC) presents the first advanced mobile 
telephone based on code that is completely open, in the Open Source in Mobile conference. It 
is called Neo1973, it costs 350 dollars and it uses a software platform called OpenMoko, based 
on the kernel of Linux 2.6, GTK+, X Windows and Matchbox. 

2007, January 

The FLOSSImpact [80] study, on the effect (especially the economic effect) of free software, is 
published. The study has been financed by the European Commission and it is the first large- 
scale study in the field. 

2007, 23 rd February 

Version 3.0 of the Creative Commons Licenses is published. 

2007, 8 th pril 

Version 4.0 of Debian GNU/Linux is released. 

GNUFDL" PID_001 48385 17 Appendixes 

3. Appendix C. GNU Public License 

Version 2, June 1991 

Copyright© 1989, 1991 Free Software Foundation, Inc. 675 Mass Ave, Cam- 
bridge, MA 02139, USA 

Literal copies of this document may be copied and distributed, but not mod- 


The licenses for most software are designed to take away your freedom to share 
and change it. By contrast, the GNU General Public License is intended to 
guarantee your freedom to share and change free software to make sure the 
software is free for all its users. This General Public License applies to most 
of the Free Software Foundation's software and to any other program whose 
authors commit to using it. (Some other Free Software Foundation software 
is covered by the GNU Lesser General Public License instead.) You can apply 
it to your programs, too. 

When we speak of free software, we are referring to freedom, not price. Our 
General Public Licenses are designed to make sure that you have the freedom 
to distribute copies of free software (and charge for this service if you wish), 
that you receive source code or can get it if you want it, that you can change 
the software or use pieces of it in new free programs; and that you know you 
can do these things. 

To protect your rights, we need to make restrictions that forbid anyone to de- 
ny you these rights or to ask you to surrender the rights. These restrictions 
translate to certain responsibilities for you if you distribute copies of the soft- 
ware, or if you modify it. 

For example, if you distribute copies of such a program, whether gratis or for 
a fee, you must give the recipients all the rights that you have. You must make 
sure that they, too, receive or can get the source code. And you must show 
them these terms so they know their rights. 

We protect your rights with two steps: 

We protect your rights with two steps: (1) copyright the software, and (2) offer 
you this license which gives you legal permission to copy, distribute and/or 
modify the software. 

GNUFDL" PID_001 48385 18 Appendixes 

Also, for each author's protection and ours, we want to make certain that ev- 
eryone understands that there is no warranty for this free software. If the soft- 
ware is modified by someone else and passed on, we want its recipients to 
know that what they have is not the original, so that any problems introduced 
by others will not reflect on the original authors' reputations. 

Finally, any free program is threatened constantly by software patents. We 
wish to avoid the danger that redistributors of a free program will individually 
obtain patent licenses, in effect making the program proprietary. To prevent 
this, we have made it clear that any patent must be licensed for everyone's 
free use or not licensed at all. 

The precise terms and conditions for copying, distribution and modification 


0) This License applies to any program or other work which contains a notice 
placed by the copyright holder saying it may be distributed under the terms of 
this General Public License. The "Program", below, refers to any such program 
or work, and a "work based on the Program" means either the Program or 
any derivative work under copyright law: that is to say, a work containing 
the Program or a portion of it, either verbatim or with modifications and/or 
translated into another language. (Hereinafter, translation is included without 
limitation in the term "modification".) Each licensee is addressed as "you". 

Activities other than copying, distribution and modification are not covered 
by this License; they are outside its scope. The act of running the Program is 
not restricted, and the output from the Program is covered only if its contents 
constitute a work based on the Program (independent of having been made 
by running the Program). Whether that is true depends on what the Program 

1) You may copy and distribute verbatim copies of the Program's source code 
as you receive it, in any medium, provided that you conspicuously and appro- 
priately publish on each copy an appropriate copyright notice and disclaimer 
of warranty; keep intact all the notices that refer to this License and to the 
absence of any warranty; and give any other recipients of the Program a copy 
of this License along with the Program. 

You may charge a fee for the physical act of transferring a copy, and you may 
at your option offer warranty protection in exchange for a fee. 

GNUFDL" PID_001 48385 19 Appendixes 

2) You may modify your copy or copies of the Program or any portion of 
it, thus forming a work based on the Program, and copy and distribute such 
modifications or work under the terms of Section 1 above, provided that you 
also meet all of these conditions: 

• a) You must cause the modified files to carry prominent notices stating 
that you changed the files and the date of any change. 

• b) You must cause any work that you distribute or publish, that in whole 
or in part contains or is derived from the Program or any part thereof, to 
be licensed as a whole at no charge to all third parties under the terms 
of this License. 

• c) If the modified program normally reads commands interactively when 
run, you must cause it, when started running for such interactive use in 
the most ordinary way, to print or display an announcement including 
an appropriate copyright notice and a notice that there is no warranty (or 
else, saying that you provide a warranty) and that users may redistribute 
the program under these conditions, and telling the user how to view a 
copy of this License. (Exception: if the program itself is interactive but 
does not normally print such an announcement, your work based on the 
program is not required to print an announcement.) 

These requirements apply to the modified work as a whole. If identifiable sec- 
tions of that work are not derived from the program, and can be reasonably 
considered independent and separate works in themselves, then this License, 
and its terms, do not apply to those sections when you distribute them as 
separate works. But when you distribute the same sections as part of a whole 
which is a work based on the program, the distribution of the whole must be 
on the terms of this License, whose permissions for other licensees extend to 
the entire whole, and thus to each and every part regardless of who wrote it. 

Thus, it is not the intent of this section to claim rights or contest your rights 
to work written entirely by you; rather, the intent is to exercise the right to 
control the distribution of derivative or collective works based on the program. 

In addition, mere aggregation of another work not based on the program with 
the program (or with a work based on the program) on a volume of a storage 
or distribution medium does not bring the other work under the scope of this 

3) You may copy and distribute the program (or a work based on it, under 
Section 2) in object code or executable form under the terms of Sections 1 and 
2 above provided that you also do one of the following: 

GNUFDL» PID_001 48385 20 Appendixes 

• a) Accompany it with the complete corresponding machine-readable 
source code, which must be distributed under the terms of Sections 1 and 
2 above on a medium customarily used for software interchange; or, 

• b) Accompany it with a written offer, valid for at least three years, to give 
any third party, for a charge no more than your cost of physically perform- 
ing source distribution, a complete machine-readable copy of the corre- 
sponding source code, to be distributed under the terms of Sections 1 and 
2 above on a medium customarily used for software interchange; or, 

The source code for a work means the preferred form of the work for making 
modifications to it. For an executable work, complete source code means all 
the source code for all modules it contains, plus any associated interface def- 
inition files, plus the scripts used to control compilation and installation of 
the executable. However, as a special exception, the source code distributed 
need not include anything that is normally distributed (in either source or 
binary form) with the major components (compiler, kernel, and so on) of the 
operating system on which the executable runs, unless that component itself 
accompanies the executable. 

If distribution of executable or object code is made by offering access to copy 
from a designated place, then offering equivalent access to copy the source 
code from the same place counts as distribution of the source code, even 
though third parties are not compelled to copy the source along with the ob- 
ject code. 

4) You may not copy, modify, sublicense, or distribute the program except as 
expressly provided under this License. Any attempt otherwise to copy, modify, 
sublicense or distribute the program is void, and will automatically terminate 
your rights under this License. However, parties who have received copies, or 
rights, from you under this License will not have their licenses terminated so 
long as such parties remain in full compliance. 

5) You are not required to accept this License, since you have not signed it. 
However, nothing else grants you permission to modify or distribute the pro- 
gram or its derivative works. These actions are prohibited by law if you do not 
accept this License. Therefore, by modifying or distributing the program (or 
any work based on the program), you indicate your acceptance of this License 
to do so, and all its terms and conditions for copying, distributing or modify- 
ing the program or works based on it. 

6) Each time you redistribute the program (or any work based on the pro- 
gram), the recipient automatically receives a license from the original licensor 
to copy, distribute or modify the program subject to these terms and condi- 

GNUFDL» PID_001 48385 21 Appendixes 

tions. You may not impose any further restrictions on the recipients' exercise 
of the rights granted herein. You are not responsible for enforcing compliance 
by third parties to this License. 

7) If, as a consequence of a court judgment or allegation of patent infringe- 
ment or for any other reason (not limited to patent issues), conditions are 
imposed on you (whether by court order, agreement or otherwise) that con- 
tradict the conditions of this License, they do not excuse you from the condi- 
tions of this License. If you cannot distribute so as to satisfy simultaneously 
your obligations under this License and any other pertinent obligations, then 
as a consequence you may not distribute the program at all. For example, if 
a patent license would not permit royalty-free redistribution of the program 
by all those who receive copies directly or indirectly through you, then the 
only way you could satisfy both it and this License would be to refrain entirely 
from distribution of the program. 

If any portion of this section is held invalid or unenforceable under any par- 
ticular circumstance, the balance of the section is intended to apply and the 
section as a whole is intended to apply in other circumstances. 

It is not the purpose of this section to induce you to infringe any patents or 
other property right claims or to contest validity of any such claims; this sec- 
tion has the sole purpose of protecting the integrity of the free software distri- 
bution system, which is implemented by public license practices. Many people 
have made generous contributions to the wide range of software distributed 
through that system in reliance on consistent application of that system; it is 
up to the author/donor to decide if he or she is willing to distribute software 
through any other system and a licensee cannot impose that choice. 

This section is intended to make thoroughly clear what is believed to be a 
consequence of the rest of this License. 

8) If the distribution and/or use of the program is restricted in certain countries 
either by patents or by copyrighted interfaces, the original copyright holder 
who places the program under this License may add an explicit geographical 
distribution limitation excluding those countries, so that distribution is per- 
mitted only in or among countries not thus excluded. In such case, this Li- 
cense incorporates the limitation as if written in the body of this License. 

9) The Free Software Foundation may publish revised and/or new versions 
of the General Public License from time to time. Such new versions will be 
similar in spirit to the present version, but may differ in detail to address new 
problems or concerns. 

Each version is given a distinguishing version number. If the program specifies 
a version number of this License which applies to it and "any later version", 
you have the option of following the terms and conditions either of that ver- 

GNUFDL" PID_001 48385 22 Appendixes 

sion or of any later version published by the Free Software Foundation. If the 
program does not specify a version number of this License, you may choose 
any version ever published by the Free Software Foundation. 

10) If you wish to incorporate parts of the program into other free programs 
whose distribution conditions are different, write to the author to ask for per- 
mission. For software which is copyrighted by the Free Software Foundation, 
write to the Free Software Foundation; we sometimes make exceptions for this. 
Our decision will be guided by the two goals of preserving the free status of 
all derivatives of our free software and of promoting the sharing and reuse of 
software generally. 





How to Apply These Terms to Your New Programs 

If you develop a new program, and you want it to be of the greatest possible 
use to the public, the best way to achieve this is to make it free software which 
everyone can redistribute and change under these terms. 

GNUFDL" PID_001 48385 23 Appendixes 

To do so, attach the following notices to the program. It is safest to attach 
them to the start of each source file to most effectively convey the exclusion of 
warranty; and each file should have at least the "copyright" line and a pointer 
to where the full notice is found. 

one line to give the program's name and an idea of what it does. Copyright (C) 
yyyy name of author 

This program is free software; you can redistribute it and/or modify it under 
the terms of the GNU General Public License as published by the Free Soft- 
ware Foundation; either version 2 of the License, or (at your option) any later 

This program is distributed in the hope that it will be useful, but WITHOUT 
ANY WARRANTY; without even the implied warranty of MERCHANTABILI- 
License for more details. 

You should have received a copy of the GNU General Public License along with 
this program; if not, write to the Free Software Foundation, Inc., 51 Franklin 
Street, Fifth Floor, Boston, MA 02110-1301, USA. 

Also add information on how to contact you by electronic and paper mail. 

If the program is interactive, make it output a short notice like this when it 
starts in an interactive mode: 

Gnomovision version 69, Copyright (C) yearname of author Gnomovision 
comes with ABSOLUTELY NO WARRANTY; for details type 'show w'. This is 
free software, and you are welcome to redistribute it under certain conditions; 
type 'show c' for details. 

The hypothetical commands 'show w' and 'show c' should show the appro- 
priate parts of the General Public License. Of course, the commands you use 
may be called something other than 'show w' and 'show c'; they could even 
be mouse-clicks or menu items -whatever suits your program. 

You should also get your employer (if you work as a programmer) or your 
school, if any, to sign a "copyright disclaimer" for the program, if necessary. 
Here is a sample; alter the names: 

Yoyodyne, Inc., hereby disclaims all copyright interest in the program 
'Gnomovision' (which makes passes at compilers) written by James Hacker. 

signature ofTy Coon, 1 April 1989 Ty Coon, President of Vice 

GNUFDL" PID_001 48385 24 Appendixes 

This General Public License does not permit incorporating your program into 
proprietary programs. If your program is a subroutine library, you may con- 
sider it more useful to permit linking proprietary applications with the library 
If this is what you want to do, use the GNU Lesser General Public License in- 
stead of this License. 

GNUFDL" PID_001 48385 25 Appendixes 

4. Appendix D. Texts of some legislative proposals and 
related documents 

Below is the literal text of some of the legislative proposals mentioned in chap- 
ter 6 and of some of the related documents. 

D.l. Draft bill brought by Laffitte, Tregouet and Cabanel (France) 

We provide below a translation of the proposed law made in October 1999 
by the French senators Pierre Laffitte, Rene Tregouet and Guy Cabanel 

D.l.l. Recitals 

(Only the paragraphs on free software are included.) 

[...] In order to guarantee the perpetuity of accessible data, facilitate its ex- 
change and ensure that citizens have free access to information, the use of this 
information by the Administration must not depend on the goodwill of soft- 
ware manufacturers. It is necessary to have free systems whose development 
may be guaranteed thanks to the manufacturers' source code being available 
to all. 

Free software is currently developing very fast. There are many IT companies 
that recognise that the future of their business is not in selling software, but 
in assisting people that use it, by providing the associated services. 

Our bill would establish that, after a transitional period defined by decree, the 
use of free software will be obligatory in all public administrations. 

Proprietary software, whose source code is not freely available, may only be 
used in specific cases, when an authorisation is provided by a free software 
agency. [...] 

D.1.2. Articles 

• Article 1. On the dematerialisation of information and data exchange be- 
tween public administrations. 

State services, local administrations and public bodies will ensure that 
their information and data are put into an electronic format, with elec- 
tronic networks, as of 1 st January 2002. 

GNUFDL» PID_001 48385 26 Appendixes 

The conditions that regulate the transition from the current paper-based 
exchange and the future exchange using electronic formats and networks 
will be specified by decree. 

• Article 2. On the dematerialisation of public market processes. 

In order to guarantee a great degree of transparency and quick access to in- 
formation for companies, all public tenders and the attached documents, 

will be published in electronic formats and networks, as of 1 st January 

2002. Likewise, all bids for public tenders must be published in electronic 

formats and networks. 

A decree will determine the mechanisms of the transition to electronic 


• Article 3. On open technologies. 

Subject to the exceptions mentioned in article 4, as of 1 st January 2002, 
State services, local administrations and public bodies may only use soft- 
ware that is free to use and modify and for which the source code is avail- 
A decree will determine the terms and conditions of the transition. 

• Article 4. On the Free Software Agency. 

A Free Software Agency will be created. It will be in charge of informing 
the State services, local administration and public bodies of the conditions 
in which this law must be applied. The Agency will determine the use of 
software licenses that are appropriate in the context established by this 

The Agency will ensure the interoperability of the free software used by 
the public administrations. 

The Agency will make an inventory, for each sector, of any fields in which 
there is no available free software, no applicable software that can be freely 
used and modified or no applicable software whose source code is avail- 
able. On the basis of this inventory, the Agency will declare the relevant 
public administrations as exempt from this law. 

The Free Software Agency will be open to all Internet users, and their de- 
cisions must be preceded by consultations made on the Internet. 
A representative of the Free Software Agency will be appointed in each 

The Free Software Agency's methods of working will be established by de- 

• Article 5. On the dissemination of the modifications to the software used 
in the context of this law. 

The Free Software Agency will ensure, whilst respecting copyrights, that 
the modifications to the software are disseminated in accordance with the 
framework of this law. 

GNUFDL» PID_001 48385 27 Appendixes 

• Article 6. 

The costs incurred by the State as a result of this law will be compensated 
through increases in the rights defined in articles 575 and 5 75 A of the 
General Tax Code. 

D.2. Draft Bill of Le Deaut, Paul and Cohen (France) 

We will now provide a translation of practically the whole of the draft bill 
presented by Jean-Yves Le Deaut, Christian Paul and Pierre Cohen in April 

D.2.1. Recitals 

The tremendous growth in the use of new information technologies and 
telecommunications has made it necessary to produce accompanying legisla- 
tion. The public services and the local administrations must become the mod- 
el and engine of the information society that will guarantee individual free- 
doms, consumer safety and equal opportunities in the field in question. 

Various examples show that, despite some significant progress achieved 
thanks to the actions of the Government in the field of the information soci- 
ety, the State services tend to use communication standards that are intimate- 
ly linked to one single private provider, which means that a user or collective 
is bound to act as the client of this same provider, thereby strongly reinforcing 
the phenomena of abuse of dominant position. 

The State service often use software with source code that is not available, 
which makes it impossible to correct the bugs and faults that the suppliers 
themselves refuse to correct or check whether there are security deficiencies 
in sensitive applications. The State services use, sometimes unknowingly, soft- 
ware that secretly transmits information that is a priori considered confiden- 
tial, to foreign societies or organisations. 

However, the economic models of the software and telecommunications in- 
dustry developed by the market are based, to a large extent, on the appropri- 
ation of clientele and the exponential valuation of the obtainment of user 
profiles. These economic models reward strategies of providing incompatible 
products, of industrial secrets and of planned obsolescence and the violation 
of individual freedoms. Although the French State cannot eliminate these un- 
derlying tendencies using the law due to the transnational nature of commu- 
nication networks, it can, however, facilitate the development of an informa- 
tion society on French soil that is respectful of public freedoms, of consumer 
safety and of equal opportunities, and this would hopefully set a precedent 
for Europe and the world. 

GNUFDL" PID_001 48385 28 Appendixes 

The law is based on five principles: a citizen's right to have free access to public 
information, the perpetuity of public data, the security of the State, consumer 
safety in the information society and the principle of software interoperability 

In order to guarantee the citizen's free access to public information, the code 
of the computerised data provided by the Administration must not be linked 
to one single supplier. Open standards, in other words, those in which the 
data coding regulations are public, make it possible to guarantee free access, as 
they permit, where necessary, the development of free compatible software. 

In order to guarantee the perpetuity of the public data, the use and mainte- 
nance of the software must not depend on the goodwill of the software's cre- 
ators. It is necessary to have systems whose development is always guaranteed 
by the availability of the source code. The principle of source code availability 
in the framework of license-based contracts, which is a principle that to date 
has only been present as an option in the legislation on public utility and 
software package purchases, must become the rule and be applied to all public 
software purchases. 

We have deliberately avoided an ambiguous legislative approach based exclu- 
sively on the use of free software. It would not be appropriate for the State, re- 
gardless of the recognised quality of the free software, to favour a determined 
economic model for the publication of software. On the contrary, the obliga- 
tory resort to open communication standards and the publication of source 
code will guarantee equal opportunities, in accordance with the principles of 
interoperability of the legislation on software. 

In order to guarantee national security, it is necessary to have systems that 
are free of elements that may provide remote control of the system or the in- 
voluntary transmission of information to any third parties. We need systems 
whose source code is freely accessible to the public, so that it can be examined 
by a large number of independent world experts. The bill that we propose 
should provide more security for the State, as full working knowledge of the 
source code would eliminate the growing number of pieces of software con- 
taining "backdoors". 

The bill that we propose would likewise reinforce consumer safety in the in- 
formation society, as it would allow for the emergence of new offers of soft- 
ware without "backdoors", which would not threaten the right to a private life 
and individual freedoms. 

But for equal opportunities to emerge, it will be necessary to reaffirm and re- 
inforce the principle of interoperability in the legislation on software and leg- 
islation on compatibility. Today, both of these rights are threatened by the 
parties that benefit from their dominant monopolistic position, who put ob- 
stacles to avoid the emergence of any competition. 

GNUFDL» PID_001 48385 29 Appendixes 

In order to guarantee the interoperability of software, the intellectual or in- 
dustrial property rights of a software creator must not block the development 
of new compatible software that would compete with him. The right to com- 
patibility for all, in other words, the right to freely develop, publish and use 
original software that is compatible with other software, must be guaranteed 
by the law. Likewise, the principle of interoperability introduced by European 
laws on software must prevail over the other intellectual or industrial property 
rights that may apply. Particularly, the existence of a brand on a communica- 
tions standard or a patent on an industrial process that is necessary to imple- 
ment a communications standard, must not permit its owner to block or limit 
the free dissemination of compatible free software. 

The bill that we propose could be applied immediately. In effect, most soft- 
ware editors are prepared to adopt open communication standards, such as 
those defined in Paris, Boston and Tokyo by the World Wide Web Consortium. 
There are many proprietary software editors that are likewise prepared to pro- 
vide the French Government with the source code of their products. In addi- 
tion, the offer of free software based on the Linux operating system will cov- 
er many of the Administration's needs, now and in the future. However, the 
Administrations and its collective bodies are not sufficiently informed about 
the existence of open standards or the offers of software published with its 
source code. 

In order to facilitate the fast implementation of free standards, it is necessary 
to reinforce the role of the Inter-ministerial Commission on Technical Sup- 
port for the Development of Information Technologies and Communication 
in the Administration (Mission Interministerielle de Soutin Technique pour 
le Developpement des Technologies de l'lnformation et de la Communication 
dans lAdministration), and entrust it with the mission of carrying out and 
disseminating within the Administration, a census of the offer of open stan- 
dards and software published with its source code. If there is no market for 
this, the MTIC will be in charge of developing new standards or new software 
published with its source code. In order to carry out these new tasks, the MTIC 
will be transformed into the Agency of Information Technologies and Com- 
munication (AITC). 

When there is no market, the AITC will be in charge of developing new stan- 
dards or new software published with its source code. In order to ensure equal 
opportunities, the software developments that occur will be put in the public 
domain; therefore, these developments may be sold as proprietary software or 
as free software, according to the license freely chosen by the editor. The AITC 
will also be in charge of evaluating the levels of interoperability, perpetuity 
and security of the software purchased by the French Administration. 

More generally, the open communication systems and the availability of the 
source code are essential to guarantee the interoperability, on a European lev- 
el, between the IT systems of the different administrations and the nation- 

GNUFDL" PID_001 48385 30 Appendixes 

al public bodies, and to avoid that the interconnection between systems de- 
pend solely on the goodwill of the software editors. The AITC will also be in 
charge of participating in the international cooperation projects in the sphere 
of information technologies and communications, and of facilitating interop- 
erability with the information systems of the other European Union member 

The bill that we propose would cover the concerns listed above. It reminds us 
that the State can play an important role in the economy, preserving national 
and European interests, whilst defending the market economy. This bill would 
allow France to stand as the defender of freedom within the new information 
and communication technologies. 

D.2.2. Articles 

• Article 1. 

For all computerised data exchanges, the State Administration, the local 
administrations and the local bodies would have the obligation of using 
open communication standards, constituted by public regulations and 
procedures for exchanging digital data. 

• Article 2. 

The Administration, the public bodies and the territorial public adminis- 
trations are obliged to use software whose source code is accessible. 

• Article 3. 

All individuals or corporate entities have the right to develop, publish or 
use original software that is compatible with the communication stan- 
dards of any other software. 

• Article 4. 

A public State body will be created, called the Agency of Information Tech- 
nologies and Communications. This body would report to the Ministry of 
Industry. The AITC will have the task of reporting to and advising the State 
services, the collective bodies and the public bodies on the creation and 
identification of the technical requirements with regard to information 
and communication technologies. It will identify the needs of the pub- 
lic services with regard to equipment and software, ensure that the com- 
munication standards are harmonised and propose the technical practices 
that must be applied. It will carry out inventories in each sector of activity 
of the open standards and the available software. 

Depending on the results of the inventory, it will support the develop- 
ment of open standards and software published with its source code and 
promote the use of this type of software in the public domain to mitigate 
any deficiencies in the market. 

The AITC will favour the interoperability with the information systems of 
other EU member States and participate in the international cooperation 

GNUFDL» PID_001 48385 31 Appendixes 

projects in the sphere of information and communication technologies. 
The AITC will have a representative in each prefecture. 
The AITC's ways of working will be established by decree. 

• Article 5. 

The modes of applying this law, as well as the conditions of the transi- 
tion from the current situation, will be established by decree issued by the 
Council of State. 

• Article 6. 

The expenses incurred by the State as a result of applying this law will be 
paid using the sums established in articles 575 and 5 75 A of the General 
Tax Code. 

D.3. Bill proposed by Villanueva and Rodrich (Peru) 

We will now provide the translation into English of the literal text of most 
of Draft Bill number 2485, on the Law on Free Software in Public Agencies, 
of the Peruvian congressmen Edgar Villanueva Nunez and Jacques Rodrich 
Ackerman [223]. 

D.3.1. Recitals 

The complexity of the world we are living in demands permanent review and 
constant adaptation of its institutional framework to be up to date with the 
current technological trends that the world imposes. 

The discovery of new information technologies and among them, free soft- 
ware, has become an ideal instrument to assure the preservation of the State's 

In this way technology fulfils its role of facilitating the different and multiple 
human activities, being one of them, the handling of public information. 

According to the Peruvian Constitution, in section 5 of article 2, "all persons 
have the right to solicit information that they need without disclosing the 
reason, and to receive that information from any public entity within the pe- 
riod specified by law, at a reasonable cost. Information that affects personal 
intimacy and that is expressly excluded by law or for reasons of national se- 
curity is not subject to disclosure". 

Section 6 of the same article emphasises the right all persons have "to be as- 
sured that information services, computerised or not, public or private, do not 
provide information that affects personal and family privacy". 

GNUFDL" PID_001 48385 32 Appendixes 

Having said this, it is obvious the concern of our Constitution for establishing 
institutional bases that protect the citizens' freedom to information access and 
the non-disclosure of information that affects personal and familiar intimacy, 
likewise for reasons of national security. 

The guarantee of these rights in our Constitution isn't solely based in the good- 
will of the State's agents to fulfil the norms of the Constitution, but also by 
the use of technologies that in some cases contribute and in others do not, to 
an effective protection of said citizens' rights. 

It is in this context that it is of utmost importance for the State the incorpo- 
ration of those technologies that help to reinforce the exercise of the citizens' 
access to information and its due reserve in cases that require so. 

The use of free software in all of the State's agencies points in this direction. 
Basically we can say that the fundamental principles that drive the present 
Bill are tightly related to the basic guarantees of a democratic State and we 
can sum them up in the following: 

1) Free access of the citizens to public information 

2) Perpetuity of public data 

3) Security of the State and of the citizens 

To guarantee the citizens' free access to public information, it is essential that 
the coding of the data is not tied to a sole provider. The use of standard and 
open formats assures this free access, making possible the creation of compat- 
ible software. 

To guarantee the perpetuity of public data, it is indispensable that the use and 
maintenance of software do not depend on the goodwill of the providers, nor 
on monopoly conditions, imposed by those. Systems whose evolution can be 
guaranteed by the availableness of source code are needed. 

To guarantee national security it's vital to have systems that are devoid of ele- 
ments that allow remote control or the transmission of non-desired informa- 
tion to third-parties. Therefore, it is essential to have systems whose source 
code is freely accessible to the public, so that its inspection is allowed to the 
State, the citizens and a great number of independent experts in the world. 

This proposal provides more security, because the knowledge of the source 
code will eliminate the growing number of programs with spy-ware. 

GNUFDL» PID_001 48385 33 Appendixes 

In the same way, this Bill furthers the security of the citizens, both in their 
condition of legitimate holders of the information handled by the State as in 
their condition of consumers. In this last case it would allow the growth of 
an extensive supply of free software devoid of potential spy-ware that makes 
it possible to jeopardise private life and individual freedoms. 

The State, looking to improve the quality of public administration as both 
keeper and manager of private information, will establish the conditions in 
which agencies of the State will acquire software in the future, that is, in a 
manner that is compatible with the constitutional guarantees and basic prin- 
ciples previously stated. 

The project clearly states that any given software in order to be acceptable 
for the State must not only be technically adequate to carry out a given task, 
but must also fulfil some requirements in license matters, without which the 
State could not guarantee the citizens the adequate process of their data, look- 
ing over for their integrity, confidentiality and permanent accessibility, all of 
which are critical elements for its fulfilment. 

The State establishes conditions for the use of software by the agencies of the 
State, without meddling in any way in the transactions of the private sector. 
It is acknowledged that the State does not have the ample spectrum of con- 
tractual freedom that the private sector has, because it is restricted due to the 
requirement of transparency of all public acts, and in this sense the common 
benefit must be the leading factor to take into account when legislating over 
this matter. 

The project also guarantees the principle of equality before the Law, because 
no natural or legal entity is excluded of the right to purvey those goods, under 
the conditions stated in this Bill and without any more limitations than the 
ones that are established in the Bill of Contracts and Acquisitions of the State 
(TUO Supreme Decree number 01 2- 2001 -PCM). 

Additionally to these advantages we could highlight benefits that would be- 
gin to show up as a consequence of these measures, immediately after being 
carried out. 

To begin with, there are the job opportunities for local programmers. Of the 
universe of server Software commercialised in the USA. over the last year, 27% 
belongs to "free" software, a truly significant portion for that huge and com- 
petitive market. The number speaks for itself and constitutes a firm answer to 
those who would think that free software would imply a hefty limitation to 
the employment of programmers of the country. On the contrary, the initia- 
tive will allow the release of a great amount of resources, and an incentive to 
boost human creativity. 

GNUFDL» PID_001 48385 34 Appendixes 

By making use of free software, professionals can analyse the root of the prob- 
lems and improve the development in whatever cases are necessary, using the 
globally available free software, under different licenses. It is an ideal area to 
employ creativity, an aspect in which young Peruvians would be able to reach 
good levels. 

On the other hand, by means of the free software we get rid of illegal software 
that is present in some agencies of the State. The non-permitted use of software 
inside the State or the mere suspicion of this constitutes a powerful incentive 
to make any given public employee modify the situation that goes against 
intellectual property. 

Although it is correct to say that the adoption of free software is not necessary 
to abide by the law, its use will drastically reduce the irregular occurrences and 
will act as a medium of legal infection, both in the State and the private sector. 

We can count many countries that are formally acknowledging an exclusive 
use of Free Software in the public sector. 

Among them we have France, where a legal norm about this subject is being 
debated. The government of the city of Mexico (DF) has already started an 
important migration to adopt free software in a general way and this is the 
leading country in this field in the western world. Also, in Brazil, the State of 
Recife has ruled its adoption. The Popular Republic of China has been using 
free software for several years as a policy of the State. The same applies to 
Scandinavian countries. In the US both NASA and the US Navy among other 
organisations have adopted free software for some of their needs, as have also 
done so other government and private entities. 

Finally, the project grants the execution of this law to the Presidency of the 
Council of Ministers for being this organism the one that concentrates the 
direction of all government institutions. In this sense it has a strategic advan- 
tage for carrying out the given reform and the migratory process of proprietary 
software to free software. 

These are the types of ideas in which these aspects have been specified in this 
legislative proposal. 

D.3.2. Cost/Benefit analysis 

This initiative does not imply any expense to the national treasury. However, 
for the fulfilment of its aims, it will be necessary to reassign the governmental 
expenditure whose incidence confines itself to what is effectively expended 
by each governmental organism in the processes of contracts and bids of the 
State for the acquisition of software. 

GNUFDL" PID_001 48385 35 Appendixes 

Although it is true that free software represents a substantial saving for the 
State's economy, when compared with proprietary software, this is not the 
central point of support of this Bill. As we have pointed out, its advantage 
focuses on the technological reassurances that the program conveys to the 
information that the State handles, information that in many instances is of 
a reserved nature. 

In this sense a better protection of the citizens' rights constitutes a non-mea- 
surable benefit that must be taking into account from the cost/benefit analysis 
point of view. 

We can sum up the benefits of the project in the following subjects: 

• National Security. 

In order to perform its functions, the State must store and process infor- 
mation on its citizens. The relationship between the individual and the 
State depends on the privacy and integrity of this data, which must be 
adequately kept against three specific risks: 

1) Disclosure risk: confidential data must be handled in such way that the 
access to them is made possible only to authorised persons and institu- 

2) Risk of impossibility of access: the data must be stored in such way that 
the access to them by authorised persons and institutions is guaranteed 
for all its period of usefulness. 

3) Risk of alteration: the alteration of data must be restricted, again only 
to those authorised to do so. 

With free software all these risks are considerably mitigated. It allows the 
user to make a complete and exhaustive inspection of the mechanisms 
that are used to process data. The fact that free software allows the in- 
spection of its sources is an excellent security measure because having the 
mechanisms exposed to the eyes of trained professionals makes hiding 
malicious functions inside them exponentially more difficult, even if the 
end user does not take the time to search for them by himself. 

• Technological Independence. 

With proprietary software there is no freedom of contract in the aspects of 
extension and correction of the system in use, a technological dependence 
is forged, one in which the provider is in the position of ruling, one-way 
only, terms, deadlines and costs. 

Free Software entitles the users with the freedom to control, correct and 
modify the program to suit it better to their needs. This freedom is not 
aimed at programmers only. Although they are the ones who can take 
advantage of it first-hand, the users benefit greatly too, for in this way they 
can hire any programmer (not necessarily the original author) to correct 
given errors or add functionality. 

• Local development. 

GNUFDL" PID_001 48385 36 Appendixes 

In the case of proprietary software, the user is able to execute or run a 
program, but not to inspect or modify it; consequently, the user cannot 
learn from it; the users become dependent on a technology that not on- 
ly do they not understand but that is expressly prohibited to them. The 
professionals in their environment, who could help the users to achieve 
their aims, are equally limited: as the way in which the program works 
is secret and its inspection is not permitted, it is not possible to fix it. In 
this way, local professionals see their possibilities of offering added value 
constantly more limited and their employment horizons narrow, along 
with their chances to learn more. With free software, these disadvantages 
of proprietary software are enormously mitigated. 

• Cost of software. 

The cost is greatly reduced because, being free, there is no need to ask for 
additional licenses to continue using the program. This need does exist 
with proprietary software. It is important for the user to be able to keep 
these costs under control, because if he cannot, he might be impeded to 
further carry on with his goals, bound by unplanned occurrences. Again, 
here it is the technological dependence that threatens free software. 

• More sources of employment. 

With free software, handwork that was chained as a consequence of the 
technological dependency of the State to proprietary software is freed. 
Now user resources (in this case the State agencies) will be assigned for 
maintenance and support of free software. 

• Boost to creativity and entrepreneurship. 

D.3.2.1. Costs 

The big cost that is involved with the change from proprietary to free software 
is limited to the migratory process. Even if it is true that the migratory pro- 
cess involves costs in studies, decision making to implement the new systems, 
handwork to implement the change, data conversion, retraining of personal 
and eventually expenses in licenses and/or development and time; it is no less 
certain that all these are fixed costs and are paid only once. 

On the other hand, proprietary software has its costs, which were paid and 
which cannot be recovered. But aside from these costs there are others in- 
volved with proprietary software: permanent updates (sometimes reinforced 
by a self-supported monopoly) and above all the huge price for the State that 
is the loss of the freedoms that guarantee the control of its own information. 
These costs are permanent and with the passage of time, sooner or later they 
exceed the fixed costs of carrying out a migration. 

To summarise, the benefits of the migratory process exceed its costs. 

GNUFDL" PID_001 48385 37 Appendixes 

D.3.3. Legal Formula 

D.3.3.1. Article 1. Aim of the law 

Employ exclusively free software in all the systems and computer equipment 
of every State agency. 

D.3.3. 2. Article 2. Scope of application 

The Executive, Legislative and Judicial branches as well as the autonomous 
regional or local decentralised organisms and the corporations where the State 
holds the majority of the shares will use free software in their systems and 
computer equipment. 

D.3.3. 3. Article 3. Authority of application 

The authority in charge to execute the law shall be the Council of Ministers. 

D.3.3. 4. Article 4. Definition of free software 

For the purposes of this law, program or free software shall be defined as that 
whose license shall guarantee the user, without additional cost, the following: 

• Unrestricted use of the program for any purpose. 

• Unrestricted access to the respective source code. 

• Exhaustive inspection of the working mechanisms of the program. 

• Use of the internal mechanisms and arbitrary portions of the software, to 
adapt them to the needs of the user. 

• Freedom to make and distribute copies of the software. 

• Modification of the software and freedom to distribute said modifications 
of the new resulting software, under the same license of the original soft- 

D.3.3. 5. Article 5. Exceptions 

Given the case where no solution which uses free software exists, that could 
satisfy the determined necessity, the State Agencies could adopt the following 
alternatives adhering to their order. 

GNUFDL» PID_001 48385 38 Appendixes 

If verifiable time constraints should occur in attending a technical problem 
and proprietary software was found to be available, the organism that needed 
it could negotiate a permission of exception before the competent authority 
to utilise proprietary software that has the following characteristics: 

• The programs shall comply with the stipulations mentioned in section 4 
of the law, except for the free distribution of the modified program. In 
such a case the permission of exception could be definitive. 

• If no programs of the preceding category were available, those that exist in 
a free project of advanced development shall be chosen. The permission 
in this case shall be transitory and will automatically expire when the free 
software becomes mature with the functionality that is necessary. 

• If no products could be found that met these conditions, then proprietary 
software could be used, but the demanded permission of exception from 
the competent authority will expire automatically two years after it was 
issued, having to be renewed previous establishment that a satisfactory 
solution of free software was not available. 

The competent authority shall emit a permission of exception only if the State 
organism guarantees the storage of data in open formats, without prejudice 
of payment for the proprietary licenses. 

D.3.3.6. Article 6. Educational permissions 

All educational establishments that depend on the State are able to manage its 
proprietary software license of its use in research, after paying the correspond- 
ing intellectual property rights and applicable licenses, provided that the aim 
of the research is directly associated to the use of the program in question. 

D.3.3.7. Article 7. Transparency of the exceptions 

The exceptions that originate in the authority of a given application must be 
sustained and published in the website of the State's Portal. 

The resolution that authorises the exception must enumerate the functional 
requirements that the program must fulfil. 

D.3.3.8. Article 8. Exceptional authorisation 

In case some State agency cannot fulfil its requirements with software stated 
in article 2 of this law then it is authorised to acquire proprietary software to 
store or process data which must be kept in reserve, the respective authority 
must publish in the State's portal a report where the risks associated with the 
use of given software for a particular application must be explained. 

GNUFDL» PID_001 48385 39 Appendixes 

The exceptional permissions granted to State agencies related with security 
and national defence are exempted from the previously stated obligation. 

D.3.3.9. Article 9. Responsibilities 

The maximum administrative authority and the technical and informative 
authority of each agencies of the State assume the responsibility for the fulfil- 
ment of this law. 

D.3.3.10. Article 10. Regulatory norm 

The executive branch of the government will rule within one hundred and 
eighty days deadline, the conditions, deadlines and forms in which the current 
status quo will be changed to one which satisfies the conditions of this law, 
and will guide, in that sense, all future contracts and negotiations for software 

In the same way, it will direct the migratory process of the proprietary software 
systems to free software ones, in every case where the given circumstances so 

D. 3. 3.11. Article 11. Glossary of terms 

a) Program or software: any sequence of instructions used by a digital data 
processing system to carry out a specific task or to solve a given problem. 

b) Execution or use of a program: the act of using it on any digital data pro- 
cessing system to carry out a function. 

c) User: natural or legal entity that makes use of the software. 

d) Source code or source program: complete set of instructions and source 
digital files created or modified by those who programmed them, plus all the 
support digital files, like data tables, images, specifications, documentation, 
and any other element that is necessary to create the executable program. 
As an exception, all those tools that are usually available as free software in 
other media may be excluded, for example, compilers, operating systems and 

e) Free software or program: that which guarantees the user, without further 
cost, the following: 

• Unrestricted use of the program for any purpose. 

• Unrestricted access to the respective source code. 

GNUFDL" PID_001 48385 40 Appendixes 

• Exhaustive inspection of the working mechanisms of the program. 

• Use of the internal mechanisms and arbitrary portions of the software, to 
adapt them to the needs of the user. 

• Freedom to make and distribute copies of the software. 

• Modification of the software and freedom to distribute said modifications 
of the new resulting software, under the same license of the original soft- 

f) Proprietary software (non-free software), that which does not fulfil all the 
requirements listed in the previous statement. 

g) Open format: any manner of digitally coded information that satisfies both 
existent standards and the following conditions: 

• Its technical documentation is publicly available. 

• The source code of at least one complete reference implementation is pub- 
licly available. 

• There are no restrictions for the creation of programs that store, transmit, 
receive or access data codified in such way. 

D.4. Letters from Microsoft Peru and congressman Villanueva 

On 21 st March 2002, Juan Alberto Gonzalez, the general manager of Microsoft 
Peru, sent a letter to congressman Edgar Villanueva Nunez with regard to his 

draft bill on free software [129]. On 8* April, the congressman replied [179]. 
We include here the English translation of a literal transcription of almost the 
whole text of both letters (the paragraphs not related to the draft bill have 
been edited out). 

D.4.1. Letter from Microsoft Peru 

As we arranged in our meeting, we attended the forum organised in the 

Congress of the Republic on March 6 th regarding the law that you have pro- 
posed. There we got the chance to listen to several presentations on the sub- 
ject. We would now like to present our position so that you have a better view 
of the real situation. 

GNUFDL» PID_001 48385 41 Appendixes 

Your proposal mandates that every public organisation exclusively uses free 
software, also known as open source software. This is something which trans- 
gresses the principles of equality before the law, of no discrimination, of free 
private initiative, and of freedom of industry and contracting, which are pro- 
tected by the Constitution. 

Your proposal, by making mandatory the use of open source software, estab- 
lishes discriminatory and non-competitive treatment in contracting and ac- 
quisitions by public organisations, violating the basic principles of the Law of 
State Contracting and Acquisitions (Number 26850). 

By forcing the State to favour a business model supporting exclusively open 
source software, your proposal will discourage local and international software 
manufacturers who make real and important investments in the country, cre- 
ate a significant number of direct and indirect jobs, and thus contribute to the 
national income. In contrast, open source software development always has a 
lesser benefit to the economy, since it mainly creates jobs in the service sector. 

Your proposal imposes the use of open source software without considering 
the risks this carries to security, warranty, and possible violation of the intel- 
lectual property rights of third parties. 

It erroneously assumes that open source software is free software, that is, without 
cost, and therefore arrives at incorrect conclusions about money saved by the 
State. It has no cost-benefit analysis to back up this assumption. 

It is wrong to think that open source software is free. Research by the Gartner 
Group (an important market researcher in the technology world, well-known 
worldwide) has shown that the cost of software acquisition (operating system 
and applications) is only 8% of the total cost of ownership that enterprises 
and organisations must face as a consequence of the rational and productive 
use of technology. The other 92% is costs of installation, training, support, 
maintenance, management, and downtime. 

One of the arguments supporting your proposal is the supposed cheapness of 
open source software when compared to commercial software, without con- 
sidering the possibility of volume licensing models. The State can really ben- 
efit from these, as other countries have. 

Additionally, the approach chosen by your project (i) is clearly more expen- 
sive because of the high costs of migration; (ii) risks loss of interoperability 
among information systems, both inside the State and between the State and 
the private sector, due to the many different distributions of open source soft- 
ware on the market. 

GNUFDL» PID_001 48385 42 Appendixes 

In most cases, open source software does not offer adequate levels of service 
to achieve better productivity by its users, nor does it offer warranties from 
well-known manufacturers. These things have caused many public entities to 
go back on their decisions to use open source software; they are now using 
commercial software in its place. 

This project discourages creativity in the Peruvian software industry, which 
sells USD 40 million worth of goods every year, USD 4 million of that export- 
ed (10 th place in the ranking of Peruvian non-traditional exports, more than 
handcrafted goods) and is a source of highly skilled jobs. With a law encourag- 
ing the use of open source software, programmers lose their intellectual prop- 
erty rights and their most important source of remuneration. 

Since open source software can be freely distributed, it cannot make any mon- 
ey for its developers by exportation. In this way, it weakens the multiplier ef- 
fect of software sales to other countries and stunts the growth of this local 
industry, which the State should be stimulating. 

In the forum, the importance of the use of open source software in education 
was discussed, without commenting on the complete failure of this initiative 
in countries like Mexico. There, the same State officials who supported the 
project now say that open source software did not provide a learning experi- 
ence to children in schools, adequate levels of training were not available na- 
tionwide, inadequate support for the platform was provided, and the software 
was not integrated well enough with existing school computer systems. 

If open source software fulfils all the requirements of State entities, why should 
a law be needed to adopt its use? Should not the market freely choose which 
products provide more benefits and value? 

D.4.2. Reply from Congressman Villanueva 

First of all, I thank you for your letter of March 25, 2002 in which you state 
the official position of Microsoft relative to Bill Number 1609, Free Software 
in Public Administration, which is indubitably inspired by the desire for Peru 
to find a suitable place in the global technological context. In the same spirit, 
and convinced that we will find the best solutions through an exchange of 
clear and open ideas, I will take this opportunity to reply to the commentaries 
included in your letter. 

While acknowledging that opinions such as yours constitute a significant con- 
tribution, it would have been even more worthwhile for me if, rather than 
formulating objections of a general nature (which we will analyse in detail 
later) you had gathered solid arguments for the advantages that proprietary 

GNUFDL" PID_001 48385 43 Appendixes 

software could bring to the Peruvian State, and to its citizens in general, since 
this would have allowed a more enlightening exchange in respect of each of 
our positions. 

With the aim of creating an orderly debate, we will assume that what you call 
open source software is what the bill defines as free software, since there exists 
software for which the source code is distributed together with the program, 
but which does not fall within the definition established by the bill; and that 
what you call commercial software is what the bill defines as proprietary or non- 
free, given that there exists free software which is sold in the market for a price 
like any other good or service. 

It is also necessary to make it clear that the aim of the bill we are discussing is 
not directly related to the amount of direct savings that can be made by using 
free software in state institutions. That is in any case a marginal aggregate 
value, but in no way is it the chief focus of the bill. The basic principles which 
inspire the bill are linked to the basic guarantees of a state of law, such as: 

• free access to public information by the citizen, 

• permanence of public data, 

• security of the State and citizens. 

To guarantee the citizens' free access to public information, it is essential that 
the coding of the data is not tied to a sole provider. The use of standard and 
open formats gives a guarantee of this free access, if necessary through the 
creation of compatible free software. 

To guarantee the permanence of public data, it is necessary that the usability 
and maintenance of the software does not depend on the goodwill of the 
suppliers, or on the monopoly conditions imposed by them. For this reason 
the State needs systems the development of which can be guaranteed due to 
the availability of the source code. 

To guarantee national security or the security of the State, it is indispensable 
to be able to rely on systems without elements which allow control from a 
distance or the undesired transmission of information to third parties. There- 
fore, systems with source code freely accessible to the public are required to 
allow their inspection by the State itself, by the citizens, and by a large num- 
ber of independent experts throughout the world. Our proposal brings further 
security, since the knowledge of the source code will eliminate the growing 
number of programs with spy code. 

GNUFDL" PID_001 48385 44 Appendixes 

In the same way, our proposal strengthens the security of the citizens, both 
in their role as legitimate owners of information managed by the State, and 
in their role as consumers; in this second case, by allowing the growth of a 
widespread availability of free software not containing spy code able to put at 
risk privacy and individual freedoms. 

In this sense, the bill is limited to establishing the conditions under which 
the State bodies will obtain software in the future, that is, in a way compatible 
with these basic principles. 

From reading the bill it will be clear that once passed: 

the law does not forbid the production of proprietary software 

the law does not forbid the sale of proprietary software 

the law does not specify which concrete software to use 

the law does not dictate the supplier from whom software will be bought 

the law does not limit the terms under which a software product can be 

What the bill does express clearly, is that, for software to be acceptable for 
the State it is not enough that it is technically capable of fulfilling a task, but 
that further the contractual conditions must satisfy a series of requirements 
regarding the license, without which the State cannot guarantee the citizen 
adequate processing of his data, watching over its integrity, confidentiality, 
and accessibility throughout time, as these are very critical aspects for its nor- 
mal functioning. 

We agree, Mr. Gonzalez, that information and communication technology 
have a significant impact on the quality of life of the citizens (whether it be 
positive or negative). We surely also agree that the basic values I have pointed 
out above are fundamental in a democratic state like Peru. So we are very 
interested to know of any other way of guaranteeing these principles, other 
than through the use of free software in the terms defined by the bill. 

As for the observations you have made, we will now go on to analyse them 
in detail: 

Firstly, you point out that: 1. "Your proposal mandates that every public or- 
ganisation exclusively use free software, also known as open source software. 
This is something which transgresses the principles of equality before the law, 
of no discrimination, of free private initiative, and of freedom of industry and 
contracting, which are protected by the Constitution." 

GNUFDL" PID_001 48385 45 Appendixes 

This understanding is in error. The bill in no way affects the rights you list; 
it limits itself entirely to establishing conditions for the use of software on 
the part of state institutions, without in any way meddling in private sector 
transactions. It is a well established principle that the State does not enjoy the 
wide spectrum of contractual freedom of the private sector, as it is limited in 
its actions precisely by the requirement for transparency of public acts; and 
in this sense, the preservation of the greater common interest must prevail 
when legislating on the matter. 

The bill protects equality under the law, since no natural or legal entity is 
excluded from the right of offering these goods to the State under the condi- 
tions defined in the bill and without more limitations than those established 
by the Law of State Contracts and Purchasing (TUO by Supreme Decree No. 

The bill does not introduce any discrimination whatsoever, since it only estab- 
lishes how the goods have to be provided (which is a State power) and not who 
has to provide them (which would effectively be discriminatory, if restrictions 
based on national origin, race religion, ideology, sexual preference etc. were 
imposed). On the contrary, the bill is decidedly anti-discriminatory This is so 
because by defining with no room for doubt the conditions for the provision 
of software, it prevents State bodies from using software which has a license 
including discriminatory conditions. 

It should be obvious from the preceding two paragraphs that the bill does not 
harm free private enterprise, since the latter can always choose under what 
conditions it will produce software; some of these will be acceptable to the 
State, and others will not be since they contradict the guarantee of the basic 
principles listed above. This free initiative is of course compatible with the 
freedom of industry and freedom of contract (in the limited form in which the 
State can exercise the latter). Any private subject can produce software under 
the conditions which the State requires, or can refrain from doing so. Nobody 
is forced to adopt a model of production, but if they wish to provide software 
to the State, they must provide the mechanisms which guarantee the basic 
principles, and which are those described in the bill. 

By way of an example: nothing in the text of the bill would prevent your com- 
pany offering the State bodies an office suite, under the conditions defined 
in the bill and setting the price that you consider satisfactory. If you did not, 
it would not be due to restrictions imposed by the law, but to business deci- 
sions relative to the method of commercialising your products, decisions with 
which the State is not involved. 

To continue, you note that: 2. "Your proposal, by making mandatory the use of 
open source software, establishes discriminatory and non-competitive treat- 
ment in contracting and acquisitions by public organizations...". 

GNUFDL" PID_001 48385 46 Appendixes 

This statement is just a reiteration of the previous one, and so the response 
can be found above. However, let us concern ourselves for a moment with 
your comment regarding "non-competitive practices." 

Of course, in defining any kind of purchase, the buyer sets conditions which 
relate to the proposed use of the good or service. From the start, this excludes 
certain manufacturers from the possibility of competing, but does not exclude 
them a priori, but rather based on a series of principles determined by the au- 
tonomous will of the purchaser, and so the process takes place in conformance 
with the law. And in the bill it is established that no-one is excluded from com- 
peting as far as he guarantees the fulfilment of the basic principles. 

Furthermore, the bill stimulates competition, since it tends to generate a sup- 
ply of software with better conditions of usability, and to better existing work, 
in a model of continuous improvement. 

On the other hand, the central aspect of competition is the chance to provide 
better choices to the consumer. Now, it is impossible to ignore the fact that 
marketing does not play a neutral role when the product is offered on the 
market (since accepting the opposite would lead one to suppose that firms' 
expenses in marketing lack any sense), and that therefore a significant expense 
under this heading can influence the decisions of the purchaser. This influence 
of marketing is in large measure reduced by the bill that we are backing, since 
the choice within the framework proposed is based on the technical merits of 
the product and not on the effort put into commercialisation by the producer; 
in this sense, competition is increased, since the smallest software producer 
can compete on equal terms with the most powerful corporations. 

It is necessary to stress that there is no position more anti-competitive than 
that of the big software producers, which frequently abuse their dominant 
position, since in innumerable cases they propose as a solution to problems 
raised by users: "update your software to the new version" (at the user's ex- 
pense, naturally); furthermore, it is common to find arbitrary cessation of 
technical help for products, which, in the provider's judgement alone, are 
old; and so, to receive any kind of technical assistance, the user finds himself 
forced to migrate to new versions (with non-trivial costs, especially as changes 
in hardware platform are often involved). And as the whole infrastructure is 
based on proprietary data formats, the user stays trapped in the need to con- 
tinue using products from the same supplier, or to make the huge effort to 
change to another environment (probably also proprietary). 

You add: 3. "By forcing the State to favour a business model supporting exclu- 
sively open source software, your proposal will discourage local and interna- 
tional software manufacturers who make real and important investments in 
the country create a significant number of direct and indirect jobs, and thus 

GNUFDL» PID_001 48385 47 Appendixes 

contribute to the national income. In contrast, open source software develop- 
ment always has a lesser benefit to the economy, since it mainly creates jobs 
in the service sector." 

I do not agree with your statement. Partly because of what you yourself point 
out in paragraph 6 of your letter, regarding the relative weight of services in 
the context of software use. This contradiction alone would invalidate your 
position. The service model, adopted by a large number of companies in the 
software industry, is much larger in economic terms, and with a tendency to 
increase, than the licensing of programs. 

On the other hand, the private sector of the economy has the widest possible 
freedom to choose the economic model which best suits its interests, even if 
this freedom of choice is often obscured subliminally by the disproportionate 
expenditure on marketing by the producers of proprietary software. 

In addition, a reading of your opinion would lead to the conclusion that the 
State market is crucial and essential for the proprietary software industry, to 
such a point that the choice made by the State in this bill would completely 
eliminate the market for these firms. If that is true, we can deduce that the 
State must be subsidising the proprietary software industry. In the unlikely 
event that this were true, the State would have the right to apply the subsidies 
in the area it considers of greatest social value; it is undeniable, in this improb- 
able hypothesis, that if the State decided to subsidise software, it would have 
to do so choosing the free over the proprietary, considering its social effect 
and the rational use of taxpayer's money. 

In respect of the jobs generated by proprietary software in countries like ours, 
these mainly concern technical tasks of little aggregate value; at the local lev- 
el, the technicians who provide support for proprietary software produced by 
transnational companies do not have the possibility of fixing bugs, not nec- 
essarily for lack of technical capability or of talent, but because they do not 
have access to the source code to fix it. With free software one creates more 
technically qualified employment and a framework of free competence where 
success is only tied to the ability to offer good technical support and quality 
of service, one stimulates the market, and one increases the shared fund of 
knowledge, opening up alternatives to generate services of greater total value 
and a higher quality level, to the benefit of all involved: producers, service 
organisations, and consumers. 

It is a common phenomenon in developing countries that local software in- 
dustries obtain the majority of their takings in the service sector, or in the 
creation of ad hoc software. Therefore, any negative impact that the applica- 
tion of the bill might have in this sector will be more than compensated by a 
growth in demand for services (as long as these are carried out to high quali- 
ty standards). If the transnational software companies decide not to compete 
under these new rules of the game, it is likely that they will undergo some 

GNUFDL» PID_001 48385 48 Appendixes 

decrease in takings in terms of payment for licences; however, considering 
that these firms continue to allege that much of the software used by the State 
has been illegally copied, one can see that the impact will not be very seri- 
ous. Certainly, in any case their fortune will be determined by market laws, 
changes in which cannot be avoided; many firms traditionally associated with 
proprietary software have already set out on the road (supported by copious 
expense) of providing services associated with free software, which shows that 
the models are not mutually exclusive. 

With this bill the State is deciding that it needs to preserve certain fundamen- 
tal values. And it is deciding this based on its sovereign power, without af- 
fecting any of the constitutional guarantees. If these values could be guaran- 
teed without having to choose a particular economic model, the effects of the 
law would be even more beneficial. In any case, it should be clear that the 
State does not choose an economic model; if it happens that there only exists 
one economic model capable of providing software which provides the basic 
guarantee of these principles, this is because of historical circumstances, not 
because of an arbitrary choice of a given model. 

Your letter continues: "4. "Your proposal imposes the use of open source soft- 
ware without considering the risks this carries to security, warranty, and pos- 
sible violation of the intellectual property rights of third parties." 

Alluding in an abstract way to "the risks this carries ", without specifically 
mentioning a single one of these supposed dangers, shows at the least some 
lack of knowledge of the topic. So, allow me to enlighten you on these points. 

On security: 

National security has already been mentioned in general terms in the initial 
discussion of the basic principles of the bill. In more specific terms, relative to 
the security of the software itself, it is well known that all software (whether 
proprietary or free) contains errors or bugs (in programmers' slang). But it is 
also well-known that the bugs in free software are fewer, and are fixed much 
more quickly, than in proprietary software. It is not in vain that numerous 
public bodies responsible for the IT security of State systems in developed 
countries require the use of free software for the same conditions of security 
and efficiency. 

What is impossible to prove is that proprietary software is more secure than 
free, without the public and open inspection of the scientific community and 
users in general. This demonstration is impossible because the model of pro- 
prietary software itself prevents this analysis, so that any guarantee of security 
is based only on promises of good intentions (biased, by any reckoning) made 
by the producer itself, or its contractors. 

GNUFDL" PID_001 48385 49 Appendixes 

It should be remembered that in many cases, the licensing conditions include 
de non-disclosure clauses which prevent the user from publicly revealing secu- 
rity flaws found in the licensed proprietary product. 

In respect of the guarantee: 

As you know perfectly well, or could find out by reading the End User License 
Agreement of the products you license, in the great majority of cases the guar- 
antees are limited to replacement of the storage medium in case of defects, 
but in no case is compensation given for direct or indirect damages, loss of 
profits, etc... If as a result of a security bug in one of your products, not fixed 
in time by yourselves, an attacker managed to compromise crucial State sys- 
tems, what guarantees, reparations and compensation would your company 
make in accordance with your licensing conditions? The guarantees of propri- 
etary software, inasmuch as programs are delivered as is, that is, in the state 
in which they are, with no additional responsibility of the provider in respect 
of function, in no way differ from those normal with free software. 

On Intellectual Property: 

Questions of intellectual property fall outside the scope of this bill, since they 
are covered by other specific laws. The free software model in no way implies 
ignorance of these laws, and in fact the great majority of free software is cov- 
ered by copyright. In reality, the inclusion of this question in your observa- 
tions shows your confusion in respect of the legal framework in which free 
software is developed. The inclusion of the intellectual property of others in 
works claimed as one's own is not a practice that has been noted in the free 
software community; whereas, unfortunately, it has been in the area of pro- 
prietary software. As an example, the condemnation by the Commercial Court 

of Nanterre, France, on 27* September 2001 of Microsoft Corp. to a penalty 
of 3 million francs in damages and interest, for violation of intellectual prop- 
erty (piracy, to use the unfortunate term that your firm commonly uses in its 

You go on to say that: 5. "It erroneously assumes that open source software is free 
software, that is, without cost, and therefore arrives at incorrect conclusions 
about money saved by the State. It has no cost-benefit analysis to back up this 

This observation is wrong; in principle, freedom and lack of cost are orthog- 
onal concepts: there is software which is proprietary and charged for (for ex- 
ample, MS Office), software which is proprietary and free of charge (MS In- 
ternet Explorer), software which is free and charged for (Red Hat, SuSE etc., 
Gnu/Linux distributions), software which is free and not charged for (Apache, 
OpenOffice, Mozilla), and even software which can be licensed in a range of 
combinations (MySQL). 

GNUFDL» PID_001 48385 50 Appendixes 

Certainly free software is not necessarily free of charge. And the text of the bill 
does not state that it has to be so, as you will have noted after reading it. The 
definitions included in the bill state clearly what should be considered free 
software, at no point referring to freedom from charges. Although the possi- 
bility of savings in payments for proprietary software licenses are mentioned, 
the foundations of the bill clearly refer to the fundamental guarantees to be 
preserved and to the stimulus to local technological development. Given that 
a democratic state must support these principles, it has no other choice than 
to use software with publicly available source code, and to exchange informa- 
tion only in standard formats. 

If the State does not use software with these characteristics, it will be weaken- 
ing basic republican principles. Luckily, free software also implies lower total 
costs; however, even given the hypothesis (easily disproved) that it was more 
expensive than proprietary software, the simple existence of an effective free 
software tool for a particular IT function would oblige the State to use it; not 
by command of this bill, but because of the basic principles we enumerated at 
the start, and which arise from the very essence of the lawful democratic State. 

You continue: 6. "It is wrong to think that open source software is free. Re- 
search by the Gartner Group (an important market researcher in the technol- 
ogy world, well-known worldwide) has shown that the cost of software ac- 
quisition (operating system and applications) is only 8% of the total cost of 
ownership that enterprises and organisations must face as a consequence of 
the rational and productive use of technology. "The other 92% consists of: in- 
stallation, training, support, maintenance, management and administration, 
and downtime." 

This argument repeats that already given in paragraph 5 and partly contra- 
dicts paragraph 3. For the sake of brevity we refer to the comments on those 
paragraphs. However, allow me to point out that your conclusion is logically 
false: even if according to the Gartner Group the cost of software is on average 
only 8% of the total cost of use, this does not in any way deny the existence 
of software which is free of charge, that is, with a licensing cost of zero. 

In addition, in this paragraph you correctly point out that the service compo- 
nents and losses due to down time make up the largest part of the total cost 
of software use, which, as you will note, contradicts your statement regarding 
the small value of services suggested in paragraph 3. Now the use of free soft- 
ware contributes significantly to reduce the remaining life-cycle costs. This re- 
duction in the costs of installation, support etc. can be noted in several areas: 
in the first place, the competitive service model of free software, support and 
maintenance for which can be freely contracted out to a range of suppliers 
competing on the grounds of quality and low cost (this is true for installation, 
enabling, and support, and in large part for maintenance). In the second place, 
due to the reproductive characteristics of the model, maintenance carried out 
for an application is easily replicable, without incurring large costs (that is, 

GNUFDL" PID_001 48385 51 Appendixes 

without paying more than once for the same thing) since modifications, if one 
wishes, can be incorporated in the common fund of knowledge. Thirdly, the 
huge costs caused by non-functioning software (blue screens of death, malicious 
code such as virus, worms, and trojans, exceptions, general protection faults 
and other well-known problems) are reduced considerably by using more sta- 
ble software. And it is well-known that one of the most remarkable virtues of 
free software is its stability. 

You further state that: 7. "One of the arguments supporting your proposal is 
the supposed cheapness of open source software when compared to commer- 
cial software, without considering the possibility of volume licensing models, 
which can be highly advantageous for the State, as has happened in other 

I have already pointed out that what is in question is not the cost of the soft- 
ware but the principles of freedom of information, accessibility, and security. 
These arguments have been covered extensively in the preceding paragraphs 
to which I would refer you. 

On the other hand, there certainly exist types of volume licensing (although 
unfortunately proprietary software does not satisfy the basic principles). But 
as you correctly pointed out in the immediately preceding paragraph of your 
letter, they only manage to reduce the impact of a component which makes 
up no more than 8% of the total. 

You continue: 8. "Additionally, the approach chosen by your project (i) is clear- 
ly more expensive because of the high costs of migration; (ii) risks loss of in- 
teroperability among information systems, both inside the State and between 
the State and the public sector, due to the many different distributions of open 
source software on the market." 

Let us analyse your statement in two parts. Your first argument, that migration 
implies high costs, is in reality an argument in favour of the bill. Because the 
more time goes by, the more difficult migration to another technology will 
become; and at the same time, the security risks associated with proprietary 
software will continue to increase. In this way, the use of proprietary systems 
and formats will make the State ever more dependent on specific suppliers. 
On the contrary, once a policy of using free software has been established 
(which certainly, does imply some cost) then on the contrary migration from 
one system to another becomes very simple, since all data is stored in open 
formats. On the other hand, migration to an open software context implies 
no more costs than migration between two different proprietary software con- 
texts, which invalidates your argument completely. 

The second argument refers to "loss of interoperability among information 
systems, both inside the State and between the State and the private sector". 
This statement implies a certain lack of knowledge of the way in which free 

GNUFDL" PID_001 48385 52 Appendixes 

software is built, which does not maximise the dependence of the user on a 
particular platform, as normally happens in the realm of proprietary software. 
Even when there are multiple free software distributions, and numerous pro- 
grams which can be used for the same function, interoperability is guaranteed 
as much by the use of standard formats, as required by the bill, as by the pos- 
sibility of creating interoperable software given the availability of the source 

You then say that: 9. "In most cases, open source software does not offer ad- 
equate levels of service to achieve better productivity by its users, nor does 
it offer warranties from well-known manufacturers. These things have caused 
many public entities to go back on their decisions to use open source software; 
they are now using commercial software in its place." 

This observation is without foundation. In respect of the guarantee, your ar- 
gument was rebutted in the response to paragraph 4. In respect of support 
services, it is possible to use free software without them (just as also happens 
with proprietary software), but anyone who does need them can obtain sup- 
port separately, whether from local firms or from international corporations, 
again just as in the case of proprietary software. 

On the other hand, it would contribute greatly to our analysis if you could 
inform us about free software projects established in public bodies which have 
already been abandoned in favour of proprietary software. We know of a good 
number of cases where the opposite has taken place, but do not know of any 
where what you describe has taken place. 

You continue by observing that: 10. "This project discourages creativity in the 
Peruvian software industry, which sells USD 40 million worth of goods every 

year, USD 4 million of that exported (10* place in the ranking of Peruvian non 
traditional exports, more than handcrafted goods) and is a source of highly 
skilled jobs. With a law encouraging the use of open source software, program- 
mers lose their intellectual property rights and their most important source 
of remuneration." 

It is clear enough that nobody is forced to commercialise their code as free 
software. The only thing to take into account is that if it is not free software, 
it cannot be sold to the public sector. This is not in any case the main market 
for the national software industry. We covered some questions referring to the 
influence of the bill on the generation of employment which would be both 
highly technically qualified and in better conditions for competition above, 
so it seems unnecessary to insist on this point. 

What follows in your statement is incorrect. On the one hand, no author of 
free software loses his intellectual property rights, unless he expressly wishes to 
place his work in the public domain. The free software movement has always 
been very respectful of intellectual property, and has generated widespread 

GNUFDL" PID_001 48385 53 Appendixes 

public recognition of authors. Names like those of Richard Stallman, Linus 
Torvalds, Guido van Rossum, Larry Wall, Miguel de Icaza, Andrew Tridgell, 
Theo de Raadt, Andrea Arcangeli, Bruce Perens, Darren Reed, Alan Cox, Er- 
ic Raymond, and many others, are recognised world-wide for their contribu- 
tions to the development of software that is used today by millions of peo- 
ple throughout the world, whilst there are many material authors of excellent 
pieces of proprietary software who remain anonymous. On the other hand, to 
say that the rewards for authors rights make up the main source of payment 
of Peruvian programmers is in any case a guess, in particular since there is no 
proof to this effect, nor a demonstration of how the use of free software by 
the State would influence these payments. 

You go on to say that: 1 1 . "Since open source software can be freely distributed, 
it cannot make any money for its developers by exportation. In this way, it 
weakens the multiplier effect of software sales to other countries and stunts 
the growth of this local industry, which the State should be stimulating." 

This statement shows once again complete ignorance of the mechanisms of 
and market for free software. It tries to claim that the market of sale of non- 
exclusive rights for use (sale of licences) is the only possible one for the soft- 
ware industry, when you yourself pointed out several paragraphs above that it 
is not even the most important one. The incentives that the bill offers for the 
growth of a supply of better qualified professionals, together with the increase 
in experience that working on a large scale with free software within the State 
will bring for Peruvian technicians, will place them in a highly competitive 
position to offer their services abroad. 

You then state that: "12. In the forum, the importance of the use of open source 
software in education was discussed, without commenting on the complete 
failure of this initiative in countries like Mexico. There, the same State officials 
who supported the project now say that open source software did not provide 
a learning experience to children in schools, adequate levels of training were 
not available nationwide, inadequate support for the platform was provided, 
and the software was not integrated well enough with existing school com- 
puter systems." 

In fact Mexico has gone into reverse with the Red Escolar (Schools Network) 
project. This is due precisely to the fact that the driving forces behind the 
Mexican project used license costs as their main argument, instead of the oth- 
er reasons specified in our project, which are far more essential. Because of this 
conceptual mistake, and as a result of the lack of effective support from the 
SEP (Secretary of State for Public Education), the assumption was made that 
to implant free software in schools it would be enough to drop their software 
budget and send them a CD ROM with GNU/Linux instead. Of course this 
failed, and it could not have been otherwise, just as school laboratories fail 
when they use proprietary software and have no budget for implementation 
and maintenance. That is exactly why our bill is not limited to making the use 

GNUFDL" PID_001 48385 54 Appendixes 

of free software mandatory, but recognises the need to create a viable migra- 
tion plan, in which the State undertakes the technical transition in an orderly 
way in order to then enjoy the advantages of free software. 

You end with a rhetorical question: 13. "If open source software fulfils all the 
requirements of State entities, why should a law be needed to adopt its use? 
Should not the market freely choose which products provide more benefits 
and value?" 

We agree that in the private sector of the economy, it must be the market that 
decides which products to use, and no State interference is permissible there. 
However, in the case of the public sector, the reasoning is not the same: as 
we have already established, the State archives, handles, and transmits infor- 
mation which does not belong to it, but which is entrusted to it by citizens, 
who have no alternative under the rule of law. As a counterpart to this legal 
requirement, the State must take extreme measures to safeguard the integrity, 
confidentiality, and accessibility of this information. The use of proprietary 
software raises serious doubts as to whether these requirements can be ful- 
filled, lacks conclusive evidence in this respect, and so is not suitable for use 
in the public sector. 

The need for a law is based, firstly, on the realisation of the fundamental prin- 
ciples listed above in the specific area of software; secondly, on the fact that 
the State is not an ideal homogeneous entity, but made up of multiple bodies 
with varying degrees of autonomy in decision making. Given that it is inap- 
propriate to use proprietary software, the fact of establishing these rules in 
law will prevent the personal discretion of any State employee from putting at 
risk the information which belongs to citizens. And above all, because it con- 
stitutes an up-to-date reaffirmation in relation to the means of management 
and communication of information used today, it is based on the republican 
principle of openness to the public. 

In conformance with this universally accepted principle, the citizen has the 
right to know all information held by the State and not covered by well-found- 
ed declarations of secrecy based on law. Now, software deals with information 
and is itself information. Information in a special form, capable of being in- 
terpreted by a machine in order to execute actions, but crucial information all 
the same because the citizen has a legitimate right to know, for example, how 
his vote is computed or his taxes calculated. And for that he must have free 
access to the source code and be able to prove to his satisfaction the programs 
used for electoral computations or calculation of his taxes. 

D.5. Decree of Measures to Promote the Knowledge Society in Andalucia 

Below are some of the articles, related to free software, of the abovementioned 
Decree on Measures to Encourage the Knowledge Society in Andalucia [99]. 

GNUFDL» PID_001 48385 55 Appendixes 

• Article 11. Educational materials in computer format. 

1. All public teaching centres will have educational materials and pro- 
grams in computerised format, preferably based on free software. In any 
case, the centres will receive these formats from the Regional Government 
of Andalucia. 

2. Likewise, the teachers will receive incentives for using computerised 
curricular materials and programs or using Internet, especially with regard 
to developments made using free software. 

• Article 31. Free software. 

1. When purchasing computer equipment that will be used in public 
teaching centres for educational activities, it should be ensured that all 
the hardware is compatible with operating systems based on free software. 
Computers will come preinstalled with all the free software that is neces- 
sary for the specific purposes for which they are intended. 

2. The computer equipment that the Regional Government of Andalucia 
provides for public access to Internet will be based on free software prod- 

3. The Regional Government of Andalucia will foster the dissemination 
and the personal, domestic and educational use of free software. For these 
purposes, an online advice service will be established for the installation 
and use of these types of products. 

• Article 49. Objective. 

1. There will be subsidies for the development of innovative projects that 
facilitate the integration of IT and communications in professional and 
occupational training. 

2. These projects will follow one of the following models: 

a) Preparation of materials and contents of professional and occupational 
training for their use and dissemination by Internet, especially with regard 
to the developments made using free software. 

b) Training initiatives using innovative methods, such as long-distance 
learning and methods whereby the students only need to attend the cours- 
es personally on certain occasions. 

GNUFDL' PID 00148385 


5. Appendix E. Creative Commons' 


Version 3.0 Unported 





1. License 



1. Definitions 

a) "Adaptation" means a work based upon the Work, or upon the Work 
and other pre-existing works, such as a translation, adaptation, derivative 
work, arrangement of music or other alterations of a literary or artistic 
work, or phonogram or performance and includes cinematographic adap- 
tations or any other form in which the Work may be recast, transformed, 
or adapted including in any form recognisably derived from the original, 
except that a work that constitutes a Collection will not be considered an 
Adaptation for the purpose of this License. For the avoidance of doubt, 
where the Work is a musical work, performance or phonogram, the syn- 
chronisation of the Work in timed-relation with a moving image ("synch- 
ing") will be considered an Adaptation for the purpose of this License. 

b) "Collection" means a collection of literary or artistic works, such as ency- 
clopedias and anthologies, or performances, phonograms or broadcasts, 
or other works or subject matter other than works listed in Section 1(f) be- 

GNUFDL" PID_001 48385 57 Appendixes 

low, which, by reason of the selection and arrangement of their contents, 
constitute intellectual creations, in which the Work is included in its en- 
tirety in unmodified form along with one or more other contributions, 
each constituting separate and independent works in themselves, which 
together are assembled into a collective whole. A work that constitutes a 
Collection will not be considered an Adaptation (as defined below) for the 
purposes of this License. 

c) "Creative Commons Compatible License" means a license that is listed at that has been approved 
by Creative Commons as being essentially equivalent to this License, in- 
cluding, at a minimum, because that license: (i) contains terms that have 
the same purpose, meaning and effect as the License Elements of this Li- 
cense; and, (ii) explicitly permits the relicensing of adaptations of works 
made available under that license under this License or a Creative Com- 
mons jurisdiction license with the same License Elements as this License. 

d) "Distribute" means to make available to the public the original and copies 
of the Work or Adaptation, as appropriate, through sale or other transfer 
of ownership. 

e) "License Elements" means the following high-level license attributes as 
selected by Licensor and indicated in the title of this License: Attribution, 

f) "Licensor" means the individual, individuals, entity or entities that 
offer(s) the Work under the terms of this License. 

g) "Original Author" means, in the case of a literary or artistic work, the 
individual, individuals, entity or entities who created the Work or if no 
individual or entity can be identified, the publisher; and in addition (i) in 
the case of a performance the actors, singers, musicians, dancers, and oth- 
er persons who act, sing, deliver, declaim, play in, interpret or otherwise 
perform literary or artistic works or expressions of folklore; (ii) in the case 
of a phonogram the producer being the person or legal entity who first 
fixes the sounds of a performance or other sounds; and, (iii) in the case of 
broadcasts, the organisation that transmits the broadcast. 

h) "Work" means the literary and/or artistic work offered under the terms 
of this License including without limitation any production in the liter- 
ary, scientific and artistic domain, whatever may be the mode or form of 
its expression including digital form, such as a book, pamphlet and other 
writing; a lecture, address, sermon or other work of the same nature; a 
dramatic or dramatico-musical work; a choreographic work or entertain- 
ment in dumb show; a musical composition with or without words; a cin- 
ematographic work to which are assimilated works expressed by a process 
analogous to cinematography; a work of drawing, painting, architecture, 

GNUFDL" PID_001 48385 58 Appendixes 

sculpture, engraving or lithography; a photographic work to which are as- 
similated works expressed by a process analogous to photography; a work 
of applied art; an illustration, map, plan, sketch or three-dimensional work 
relative to geography, topography, architecture or science; a performance; 
a broadcast; a phonogram; a compilation of data to the extent it is pro- 
tected as a copyrightable work; or a work performed by a variety or circus 
performer to the extent it is not otherwise considered a literary or artistic 

i) "You" means an individual or entity exercising rights under this License 
who has not previously violated the terms of this License with respect to 
the Work, or who has received express permission from the Licensor to 
exercise rights under this License despite a previous violation. 

j) "Publicly Perform" means to perform public recitations of the Work and 
to communicate to the public those public recitations, by any means 
or process, including by wire or wireless means or public digital perfor- 
mances; to make available to the public Works in such a way that members 
of the public may access these Works from a place and at a place individ- 
ually chosen by them; to perform the Work to the public by any means or 
process and the communication to the public of the performances of the 
Work, including by public digital performance; to broadcast and rebroad- 
cast the Work by any means including signs, sounds or images. 

k) "Reproduce" means to make copies of the Work by any means including 
without limitation by sound or visual recordings and the right of fixation 
and reproducing fixations of the Work, including storage of a protected 
performance or phonogram in digital form or other electronic medium. 

2. Fair Dealing Rights. Nothing in this License is intended to reduce, limit, 
or restrict any uses free from copyright or rights arising from limitations or 
exceptions that are provided for in connection with the copyright protection 
under copyright law or other applicable laws. 

3. License Grant. Subject to the terms and conditions of this License, Licensor 
hereby grants You a worldwide, royalty-free, non-exclusive, perpetual (for the 
duration of the applicable copyright) license to exercise the rights in the Work 
as stated below: 

a) to Reproduce the Work, to incorporate the Work into one or more Collec- 
tions, and to Reproduce the Work as incorporated in the Collections; 

b) to create and Reproduce Adaptations provided that any such Adaptation, 
including any translation in any medium, takes reasonable steps to clearly 
label, demarcate or otherwise identify that changes were made to the orig- 
inal Work. For example, a translation could be marked "The original work 

GNUFDL" PID_001 48385 59 Appendixes 

was translated from English to Spanish," or a modification could indicate 
"The original work has been modified."; 

c) to Distribute and Publicly Perform the Work including as incorporated in 
Collections; and, 

d) to Distribute and Publicly Perform Adaptations. 

e) For the avoidance of doubt: 

a) Non-waivable Compulsory License Schemes. In those jurisdictions 
in which the right to collect royalties through any statutory or com- 
pulsory licensing scheme cannot be waived, the Licensor reserves the 
exclusive right to collect such royalties for any exercise by You of the 
rights granted under this License; 

b) Waivable Compulsory License Schemes. In those jurisdictions in 
which the right to collect royalties through any statutory or compul- 
sory licensing scheme can be waived, the Licensor waives the exclu- 
sive right to collect such royalties for any exercise by You of the rights 
granted under this License; and, 

c) Voluntary License Schemes. The Licensor waives the right to collect 
royalties, whether individually or, in the event that the Licensor is a 
member of a collecting society that administers voluntary licensing 
schemes, via that society, from any exercise by You of the rights grant- 
ed under this License. 

The above rights may be exercised in all media and formats whether now 
known or hereafter devised. The above rights include the right to make such 
modifications as are technically necessary to exercise the rights in other media 
and formats. Subject to Section 8(f), all rights not expressly granted by Licen- 
sor are hereby reserved. 

4. Restrictions. The license granted in Section 3 above is expressly made sub- 
ject to and limited by the following restrictions: 

a) You may Distribute or Publicly Perform the Work only under the terms of 
this License. You must include a copy of, or the Uniform Resource Iden- 
tifier (URI) for, this License with every copy of the Work You Distribute 
or Publicly Perform. You may not offer or impose any terms on the Work 
that restrict the terms of this License or the ability of the recipient of the 
Work to exercise the rights granted to that recipient under the terms of 
the License. You may not sublicense the Work. You must keep intact all 
notices that refer to this License and to the disclaimer of warranties with 
every copy of the Work You Distribute or Publicly Perform. When You Dis- 
tribute or Publicly Perform the Work, You may not impose any effective 
technological measures on the Work that restrict the ability of a recipient 

GNUFDL" PID_001 48385 60 Appendixes 

of the Work from You to exercise the rights granted to that recipient under 
the terms of the License. This Section 4(a) applies to the Work as incorpo- 
rated in a Collection, but this does not require the Collection apart from 
the Work itself to be made subject to the terms of this License. If You cre- 
ate a Collection, upon notice from any Licensor You must, to the extent 
practicable, remove from the Collection any credit as required by Section 
4(c), as requested. If You create an Adaptation, upon notice from any Li- 
censor You must, to the extent practicable, remove from the Adaptation 
any credit as required by Section 4(c), as requested. 

b) You may Distribute or Publicly Perform an Adaptation only under the 
terms of: (i) this License; (ii) a later version of this License with the same 
License Elements as this License; (iii) a Creative Commons jurisdiction 
license (either this or a later license version) that contains the same Li- 
cense Elements as this License (e.g., Attribution-ShareAlike 3.0 US)); (iv) 
a Creative Commons Compatible License. If you license the Adaptation 
under one of the licenses mentioned in (iv), you must comply with the 
terms of that license. If you license the Adaptation under the terms of 
any of the licenses mentioned in (i), (ii) or (iii) (the "Applicable License"), 
you must comply with the terms of the Applicable License generally and 
the following provisions: (I) You must include a copy of, or the URI for, 
the Applicable License with every copy of each Adaptation You Distribute 
or Publicly Perform; (II) You may not offer or impose any terms on the 
Adaptation that restrict the terms of the Applicable License or the ability 
of the recipient of the Adaptation to exercise the rights granted to that 
recipient under the terms of the Applicable License; (III) You must keep 
intact all notices that refer to the Applicable License and to the disclaimer 
of warranties with every copy of the Work as included in the Adaptation 
You Distribute or Publicly Perform; (IV) when You Distribute or Publicly 
Perform the Adaptation, You may not impose any effective technological 
measures on the Adaptation that restrict the ability of a recipient of the 
Adaptation from You to exercise the rights granted to that recipient un- 
der the terms of the Applicable License. This Section 4(b) applies to the 
Adaptation as incorporated in a Collection, but this does not require the 
Collection apart from the Adaptation itself to be made subject to the terms 
of the Applicable License. 

c) If You Distribute, or Publicly Perform the Work or any Adaptations or Col- 
lections, You must, unless a request has been made pursuant to Section 
4(a), keep intact all copyright notices for the Work and provide, reasonable 
to the medium or means You are utilising: (i) the name of the Original Au- 
thor (or pseudonym, if applicable) if supplied, and/or if the Original Au- 
thor and/or Licensor designate another party or parties (e.g., a sponsor in- 
stitute, publishing entity, journal) for attribution ("Attribution Parties") in 
Licensor's copyright notice, terms of service or by other reasonable means, 
the name of such party or parties; (ii) the title of the Work if supplied; (iii) 
to the extent reasonably practicable, the URI, if any, that Licensor specifies 

GNUFDL" PID_001 48385 61 Appendixes 

to be associated with the Work, unless such URI does not refer to the copy- 
right notice or licensing information for the Work; and (iv), consistent 
with Section 3(b), in the case of an Adaptation, a credit identifying the 
use of the Work in the Adaptation (e.g., "French translation of the Work 
by Original Author," or "Screenplay based on original Work by Original 
Author"). The credit required by this Section 4(c) may be implemented in 
any reasonable manner; provided, however, that in the case of a Adapta- 
tion or Collection, at a minimum such credit will appear, if a credit for all 
contributing authors of the Adaptation or Collection appears, then as part 
of these credits and in a manner at least as prominent as the credits for 
the other contributing authors. For the avoidance of doubt, You may only 
use the credit required by this Section for the purpose of attribution in the 
manner set out above and, by exercising Your rights under this License, 
You may not implicitly or explicitly assert or imply any connection with, 
sponsorship or endorsement by the Original Author, Licensor and/or At- 
tribution Parties, as appropriate, of You or Your use of the Work, without 
the separate, express prior written permission of the Original Author, Li- 
censor and/or Attribution Parties. 

d) Except as otherwise agreed in writing by the Licensor or as may be other- 
wise permitted by applicable law, if You Reproduce, Distribute or Publicly 
Perform the Work either by itself or as part of any Adaptations or Collec- 
tions, You must not distort, mutilate, modify or take other derogatory ac- 
tion in relation to the Work which would be prejudicial to the Original 
Author's honor or reputation. Licensor agrees that in those jurisdictions 
(e.g. Japan), in which any exercise of the right granted in Section 3(b) of 
this License (the right to make Adaptations) would be deemed to be a dis- 
tortion, mutilation, modification or other derogatory action prejudicial 
to the Original Author's honor and reputation, the Licensor will waive 
or not assert, as appropriate, this Section, to the fullest extent permitted 
by the applicable national law, to enable You to reasonably exercise Your 
right under Section 3(b) of this License (right to make Adaptations) but 
not otherwise. 

5. Representations, Warranties and Disclaimer 


GNUFDL" PID_001 48385 62 Appendixes 

6. Limitation on Liability. EXCEPT TO THE EXTENT REQUIRED BY APPLI- 

7. Termination 

a) This License and the rights granted hereunder will terminate automatical- 
ly upon any breach by You of the terms of this License. Individuals or en- 
tities who have received Adaptations or Collections from You under this 
License, however, will not have their licenses terminated provided such 
individuals or entities remain in full compliance with those licenses. Sec- 
tions 1, 2, 5, 6, 7, and 8 will survive any termination of this License. 

b) Subject to the above terms and conditions, the license granted here is per- 
petual (for the duration of the applicable copyright in the Work). Notwith- 
standing the above, Licensor reserves the right to release the Work under 
different license terms or to stop distributing the Work at any time; pro- 
vided, however that any such election will not serve to withdraw this Li- 
cense (or any other license that has been, or is required to be, granted un- 
der the terms of this License), and this License will continue in full force 
and effect unless terminated as stated above. 

8. Miscellaneous 

a) Each time You Distribute or Publicly Perform the Work or a Collection, the 
Licensor offers to the recipient a license to the Work on the same terms 
and conditions as the license granted to You under this License. 

b) Each time You Distribute or Publicly Perform an Adaptation, Licensor of- 
fers to the recipient a license to the original Work on the same terms and 
conditions as the license granted to You under this License. 

c) If any provision of this License is invalid or unenforceable under applica- 
ble law, it shall not affect the validity or enforceability of the remainder 
of the terms of this License, and without further action by the parties to 
this agreement, such provision shall be reformed to the minimum extent 
necessary to make such provision valid and enforceable. 

d) No term or provision of this License shall be deemed waived and no breach 
consented to unless such waiver or consent shall be in writing and signed 
by the party to be charged with such waiver or consent. 

e) This License constitutes the entire agreement between the parties with re- 
spect to the Work licensed here. There are no understandings, agreements 

GNUFDL» PID_001 48385 63 Appendixes 

or representations with respect to the Work not specified here. Licensor 
shall not be bound by any additional provisions that may appear in any 
communication from You. This License may not be modified without the 
mutual written agreement of the Licensor and You. 

f) The rights granted under, and the subject matter referenced, in this License 
were drafted utilising the terminology of the Berne Convention for the 
Protection of Literary and Artistic Works (as amended on September 28, 
1979), the Rome Convention of 1961, the WIPO Copyright Treaty of 1996, 
the WIPO Performances and Phonograms Treaty of 1996 and the Univer- 
sal Copyright Convention (as revised on July 24, 1971). These rights and 
subject matter take effect in the relevant jurisdiction in which the License 
terms are sought to be enforced according to the corresponding provisions 
of the implementation of those treaty provisions in the applicable nation- 
al law. If the standard suite of rights granted under applicable copyright 
law includes additional rights not granted under this License, such addi- 
tional rights are deemed to be included in the License; this License is not 
intended to restrict the license of any rights under applicable law. 

GNUFDL" PID_001 48385 64 Appendixes 

6. Appendix F. GNU Free Documentation License 

Copyright (C) 2000,2001,2002 Free Software Foundation, Inc. 51 Franklin St, 
Fifth Floor, Boston, MA 02110-1301 USA Everyone is permitted to copy and 
distribute verbatim copies of this license document, but changing it is not 


The purpose of this License is to make a manual, textbook, or other functional 
and useful document "free" in the sense of freedom: to assure everyone the 
effective freedom to copy and redistribute it, with or without modifying it, 
either commercially or noncommercially. Secondarily, this License preserves 
for the author and publisher a way to get credit for their work, while not being 
considered responsible for modifications made by others. 

This License is a kind of "copyleft", which means that derivative works of the 
document must themselves be free in the same sense. It complements the 
GNU General Public License, which is a copyleft license designed for free soft- 

We have designed this License in order to use it for manuals for free software, 
because free software needs free documentation: a free program should come 
with manuals providing the same freedoms that the software does. But this 
License is not limited to software manuals; it can be used for any textual work, 
regardless of subject matter or whether it is published as a printed book. We 
recommend this License principally for works whose purpose is instruction 
or reference. 


This License applies to any manual or other work, in any medium, that con- 
tains a notice placed by the copyright holder saying it can be distributed un- 
der the terms of this License. Such a notice grants a world-wide, royalty-free 
license, unlimited in duration, to use that work under the conditions stated 
herein. The "Document", below, refers to any such manual or work. Any mem- 
ber of the public is a licensee, and is addressed as "you". You accept the license 
if you copy, modify or distribute the work in a way requiring permission un- 
der copyright law. 

A "Modified Version" of the Document means any work containing the Doc- 
ument or a portion of it, either copied verbatim, or with modifications and/or 
translated into another language. 

GNUFDL" PID_001 48385 65 Appendixes 

A "Secondary Section" is a named appendix or a front-matter section of the 
Document that deals exclusively with the relationship of the publishers or au- 
thors of the Document to the Document's overall subject (or to related mat- 
ters) and contains nothing that could fall directly within that overall subject. 
(Thus, if the Document is in part a textbook of mathematics, a Secondary Sec- 
tion may not explain any mathematics.) The relationship could be a matter 
of historical connection with the subject or with related matters, or of legal, 
commercial, philosophical, ethical or political position regarding them. 

The "Invariant Sections" are certain Secondary Sections whose titles are desig- 
nated, as being those of Invariant Sections, in the notice that says that the 
Document is released under this License. If a section does not fit the above 
definition of Secondary then it is not allowed to be designated as Invariant. 
The Document may contain zero Invariant Sections. If the Document does 
not identify any Invariant Sections then there are none. 

The "Cover Texts" are certain short passages of text that are listed, as Front- 
Cover Texts or Back-Cover Texts, in the notice that says that the Document is 
released under this License. A Front-Cover Text may be at most 5 words, and 
a Back-Cover Text may be at most 25 words. 

A "Transparent" copy of the Document means a machine-readable copy, rep- 
resented in a format whose specification is available to the general public, that 
is suitable for revising the document straightforwardly with generic text edi- 
tors or (for images composed of pixels) generic paint programs or (for draw- 
ings) some widely available drawing editor, and that is suitable for input to 
text formatters or for automatic translation to a variety of formats suitable for 
input to text formatters. A copy made in an otherwise Transparent file format 
whose markup, or absence of markup, has been arranged to thwart or discour- 
age subsequent modification by readers is not Transparent. An image format 
is not Transparent if used for any substantial amount of text. A copy that is 
not "Transparent" is called "Opaque". 

Examples of suitable formats for Transparent copies include plain ASCII with- 
out markup, Texinfo input format, LaTeX input format, SGML or XML using 
a publicly available DTD, and standard-conforming simple HTML, PostScript 
or PDF designed for human modification. Examples of transparent image for- 
mats include PNG, XCF and JPG. Opaque formats include proprietary formats 
that can be read and edited only by proprietary word processors, SGML or 
XML for which the DTD and/or processing tools are not generally available, 
and the machine-generated HTML, PostScript or PDF produced by some word 
processors for output purposes only. 

GNUFDL» PID_001 48385 66 Appendixes 

The "Title Page" means, for a printed book, the title page itself, plus such fol- 
lowing pages as are needed to hold, legibly, the material this License requires 
to appear in the title page. For works in formats which do not have any title 
page as such, "Title Page" means the text near the most prominent appearance 
of the work's title, preceding the beginning of the body of the text. 

A section "Entitled XYZ" means a named subunit of the Document whose 
title either is precisely XYZ or contains XYZ in parentheses following text 
that translates XYZ in another language. (Here XYZ stands for a specific sec- 
tion name mentioned below, such as "Acknowledgements", "Dedications", 
"Endorsements", or "History".) To "Preserve the Title" of such a section when 
you modify the Document means that it remains a section "Entitled XYZ" ac- 
cording to this definition. 

The Document may include Warranty Disclaimers next to the notice which 
states that this License applies to the Document. These Warranty Disclaimers 
are considered to be included by reference in this License, but only as regards 
disclaiming warranties: any other implication that these Warranty Disclaimers 
may have is void and has no effect on the meaning of this License. 


You may copy and distribute the Document in any medium, either commer- 
cially or noncommercially, provided that this License, the copyright notices, 
and the license notice saying this License applies to the Document are repro- 
duced in all copies, and that you add no other conditions whatsoever to those 
of this License. You may not use technical measures to obstruct or control 
the reading or further copying of the copies you make or distribute. However, 
you may accept compensation in exchange for copies. If you distribute a large 
enough number of copies you must also follow the conditions in section 3. 

You may also lend copies, under the same conditions stated above, and you 
may publicly display copies. 


If you publish printed copies (or copies in media that commonly have printed 
covers) of the Document, numbering more than 100, and the Document's 
license notice requires Cover Texts, you must enclose the copies in covers that 
carry, clearly and legibly, all these Cover Texts: Front-Cover Texts on the front 
cover, and Back-Cover Texts on the back cover. Both covers must also clearly 
and legibly identify you as the publisher of these copies. The front cover must 
present the full title with all words of the title equally prominent and visible. 
You may add other material on the covers in addition. Copying with changes 
limited to the covers, as long as they preserve the title of the Document and 
satisfy these conditions, can be treated as verbatim copying in other respects. 

GNUFDL" PID_001 48385 67 Appendixes 

If the required texts for either cover are too voluminous to fit legibly, you 
should put the first ones listed (as many as fit reasonably) on the actual cover, 
and continue the rest onto adjacent pages. 

If you publish or distribute Opaque copies of the Document numbering more 
than 100, you must either include a machine-readable Transparent copy along 
with each Opaque copy, or state in or with each Opaque copy a computer- 
network location from which the general network-using public has access to 
download using public-standard network protocols a complete Transparent 
copy of the Document, free of added material. If you use the latter option, you 
must take reasonably prudent steps, when you begin distribution of Opaque 
copies in quantity, to ensure that this Transparent copy will remain thus ac- 
cessible at the stated location until at least one year after the last time you 
distribute an Opaque copy (directly or through your agents or retailers) of that 
edition to the public. 

It is requested, but not required, that you contact the authors of the Document 
well before redistributing any large number of copies, to give them a chance 
to provide you with an updated version of the Document. 


You may copy and distribute a Modified Version of the Document under the 
conditions of sections 2 and 3 above, provided that you release the Modified 
Version under precisely this License, with the Modified Version filling the role 
of the Document, thus licensing distribution and modification of the Modi- 
fied Version to whoever possesses a copy of it. In addition, you must do these 
things in the Modified Version: 

• A. Use in the Title Page (and on the covers, if any) a title distinct from 
that of the Document, and from those of previous versions (which should, 
if there were any, be listed in the History section of the Document). You 
may use the same title as a previous version if the original publisher of 
that version gives permission. 

• B. List on the Title Page, as authors, one or more persons or entities re- 
sponsible for authorship of the modifications in the Modified Version, to- 
gether with at least five of the principal authors of the Document (all of its 
principal authors, if it has fewer than five), unless they release you from 
this requirement. 

• C. State on the Title page the name of the publisher of the Modified Ver- 
sion, as the publisher. 

• D. Preserve all the copyright notices of the Document. 

GNUFDL" PID_001 48385 68 Appendixes 

• E. Add an appropriate copyright notice for your modifications adjacent to 
the other copyright notices. 

• F. Include, immediately after the copyright notices, a license notice giving 
the public permission to use the Modified Version under the terms of this 
License, in the form shown in the Addendum below. 

• G. Preserve in that license notice the full lists of Invariant Sections and 
required Cover Texts given in the Document's license notice. 

• H. Include an unaltered copy of this License. 

• I. Preserve the section Entitled "History", Preserve its Title, and add to it 
an item stating at least the title, year, new authors, and publisher of the 
Modified Version as given on the Title Page. If there is no section Entitled 
"History" in the Document, create one stating the title, year, authors, and 
publisher of the Document as given on its Title Page, then add an item 
describing the Modified Version as stated in the previous sentence. 

• J. Preserve the network location, if any, given in the Document for public 
access to a Transparent copy of the Document, and likewise the network 
locations given in the Document for previous versions it was based on. 
These may be placed in the "History" section. You may omit a network 
location for a work that was published at least four years before the Doc- 
ument itself, or if the original publisher of the version it refers to gives 

• K. For any section Entitled "Acknowledgements" or "Dedications", Preserve 
the Title of the section, and preserve in the section all the substance and 
tone of each of the contributor acknowledgements and/ or dedications giv- 
en therein. 

• L. Preserve all the Invariant Sections of the Document, unaltered in their 
text and in their titles. Section numbers or the equivalent are not consid- 
ered part of the section titles. 

• M. Delete any section Entitled "Endorsements". Such a section may not be 
included in the Modified Version. 

• N. Do not retitle any existing section to be Entitled "Endorsements" or to 
conflict in title with any Invariant Section. 

• O. Preserve any Warranty Disclaimers. 

If the Modified Version includes new front-matter sections or appendices that 
qualify as Secondary Sections and contain no material copied from the Doc- 
ument, you may at your option designate some or all of these sections as in- 

GNUFDL" PID_001 48385 69 Appendixes 

variant. To do this, add their titles to the list of Invariant Sections in the Mod- 
ified Version's license notice. These titles must be distinct from any other sec- 
tion titles. 

You may add a section Entitled "Endorsements", provided it contains nothing 
but endorsements of your Modified Version by various parties - for example, 
statements of peer review or that the text has been approved by an organisa- 
tion as the authoritative definition of a standard. 

You may add a passage of up to five words as a Front-Cover Text, and a passage 
of up to 25 words as a Back-Cover Text, to the end of the list of Cover Texts 
in the Modified Version. Only one passage of Front-Cover Text and one of 
Back-Cover Text may be added by (or through arrangements made by) any 
one entity. If the Document already includes a cover text for the same cover, 
previously added by you or by arrangement made by the same entity you are 
acting on behalf of, you may not add another; but you may replace the old 
one, on explicit permission from the previous publisher that added the old 

The author(s) and publisher(s) of the Document do not by this License give 
permission to use their names for publicity for or to assert or imply endorse- 
ment of any Modified Version. 


You may combine the Document with other documents released under this 
License, under the terms defined in section 4 above for modified versions, 
provided that you include in the combination all of the Invariant Sections 
of all of the original documents, unmodified, and list them all as Invariant 
Sections of your combined work in its license notice, and that you preserve 
all their Warranty Disclaimers. 

The combined work need only contain one copy of this License, and multiple 
identical Invariant Sections may be replaced with a single copy. If there are 
multiple Invariant Sections with the same name but different contents, make 
the title of each such section unique by adding at the end of it, in parentheses, 
the name of the original author or publisher of that section if known, or else 
a unique number. Make the same adjustment to the section titles in the list of 
Invariant Sections in the license notice of the combined work. 

In the combination, you must combine any sections Entitled "History" in the 
various original documents, forming one section Entitled "History"; likewise 
combine any sections Entitled "Acknowledgements", and any sections Entitled 
"Dedications". You must delete all sections Entitled "Endorsements." 


GNUFDL" PID_001 48385 70 Appendixes 

You may make a collection consisting of the Document and other documents 
released under this License, and replace the individual copies of this License 
in the various documents with a single copy that is included in the collection, 
provided that you follow the rules of this License for verbatim copying of each 
of the documents in all other respects. 

You may extract a single document from such a collection, and distribute it 
individually under this License, provided you insert a copy of this License into 
the extracted document, and follow this License in all other respects regarding 
verbatim copying of that document. 


A compilation of the Document or its derivatives with other separate and in- 
dependent documents or works, in or on a volume of a storage or distribution 
medium, is called an "aggregate" if the copyright resulting from the compila- 
tion is not used to limit the legal rights of the compilation's users beyond what 
the individual works permit. When the Document is included in an aggregate, 
this License does not apply to the other works in the aggregate which are not 
themselves derivative works of the Document. 

If the Cover Text requirement of section 3 is applicable to these copies of the 
Document, then if the Document is less than one half of the entire aggregate, 
the Document's Cover Texts may be placed on covers that bracket the Docu- 
ment within the aggregate, or the electronic equivalent of covers if the Doc- 
ument is in electronic form. Otherwise they must appear on printed covers 
that bracket the whole aggregate. 


Translation is considered a kind of modification, so you may distribute trans- 
lations of the Document under the terms of section 4. Replacing Invariant Sec- 
tions with translations requires special permission from their copyright hold- 
ers, but you may include translations of some or all Invariant Sections in ad- 
dition to the original versions of these Invariant Sections. You may include a 
translation of this License, and all the license notices in the Document, and 
any Warranty Disclaimers, provided that you also include the original En- 
glish version of this License and the original versions of those notices and 
disclaimers. In case of a disagreement between the translation and the origi- 
nal version of this License or a notice or disclaimer, the original version will 

If a section in the Document is Entitled "Acknowledgements", "Dedications", 
or "History", the requirement (section 4) to Preserve its Title (section 1) will 
typically require changing the actual title. 

GNUFDL» PID_001 48385 71 Appendixes 


You may not copy, modify, sublicense, or distribute the Document except as 
expressly provided for under this License. Any other attempt to copy, modi- 
fy, sublicense or distribute the Document is void, and will automatically ter- 
minate your rights under this License. However, parties who have received 
copies, or rights, from you under this License will not have their licenses ter- 
minated so long as such parties remain in full compliance. 


The Free Software Foundation may publish new, revised versions of the GNU 
Free Documentation License from time to time. Such new versions will be 
similar in spirit to the present version, but may differ in detail to address new 
problems or concerns. See 

Each version of the License is given a distinguishing version number. If the 
Document specifies that a particular numbered version of this License "or any 
later version" applies to it, you have the option of following the terms and 
conditions either of that specified version or of any later version that has been 
published (not as a draft) by the Free Software Foundation. If the Document 
does not specify a version number of this License, you may choose any version 
ever published (not as a draft) by the Free Software Foundation. 

How to use this License for your documents 

To use this License in a document you have written, include a copy of the 
License in the document and put the following copyright and license notices 
just after the title page: 

Copyright (c) YEAR YOUR NAME. Permission is granted to copy, distribute 
and/or modify this document under the terms of the GNU Free Documenta- 
tion License, Version 1.2 or any later version published by the Free Software 
Foundation; with no Invariant Sections, no Front-Cover Texts, and no Back- 
Cover Texts. A copy of the license is included in the section entitled "GNU 
Free Documentation License". 

If you have Invariant Sections, Front-Cover Texts and Back-Cover Texts, re- 
place the "with. ..Texts." line with this: 

with the Invariant Sections being LIST THEIR TITLES, with the Front-Cover 
Texts being LIST, and with the Back-Cover Texts being LIST. 

If you have Invariant Sections without Cover Texts, or some other combina- 
tion of the three, merge those two alternatives to suit the situation. 

GNUFDL» PID_001 48385 72 Appendixes 

If your document contains nontrivial examples of program code, we recom- 
mend releasing these examples in parallel under your choice of free software 
license, such as the GNU General Public License, to permit their use in free 

Education and Culture DG 

Lifelong Learning Programme 


FrEE prnarBm LiCEncs? hduj \s FrEE 




panarBma df FrEE edftujefe"? 


in the mEDia anD in debbtee Bmans it 


ciTiEEns in BEnsraL ara ETBrTina td 


msny pedfle. Hno susn thdbe uihd ars 
FamiLiar uhth iT ara DFTEn auisrE df 


mDBTLy isnaranT aaauT DTHEra. 




Universitat Oberta 
de Catalunya