The 5 November 2004, by Enrique J. Batalla,
Maître Enrique J. Batalla, Batalla Abogados, General Secretary of the Computer Law Association (CLA)
What is free software? Does free software mean free data? Does availability of source code imply a free Internet? These are some of the questions that have lead this paper. A great number of material is provided in order for the reader and attendant to better understand the issue of free software and its implications. Free software is not a technology in itself but a licensing method product. The availability of the source code in human-readable format is the main feature of this licensing method. Nevertheless, the freedom to use and transfer such a code to third parties may be restricted depending of the kind of license involved. Therefore, the formula free software = free data may prove not to be always true.
6th Conference « Law via the Internet »
Paris, 5th November 2004
FREE SOFTWARE/FREE DATA/FREE INTERNET?
Enrique J. Batalla
Batalla Abogados, Madrid, Spain
This is quite a controversial subject that has inspired, and still inspires, “amours fous” and alarming calls of warning within the virtual and non-virtual community. “Free software” is a term that has gone beyond the world of software developers and jumped into the general public arena, partly due to the increasing dependency of our societies and productive processes on computers and the growing awareness of the citizens on this fact. The information available on the matter is everyday larger, what proves the degree of involvement of an increasing number of people regardless their background. It is actually the availability of information the key point of the debate. Information is a very, if not the most, important asset in our society. Once under the control of a few, later apparently more distributed, information seems to be going back again to the hands of a limited number of groups, in a movement regarded to by many as a reduction of the most basic freedoms.
If we understand the title of this speech as a lineal equation, free software would be then the sole foundation of a free Internet. In that case, the debate would turn out to be a struggle to prevail: either free software or a captive Internet would be the options at stake. But, let’s take the same title and include an addition symbol between the first two elements:
“Free software + free data = free Internet”, free software would be then just one of the two elements needed to achieve the goal “Free Internet”. Data could be kept free or freed by other means, not just by means of the availability of the source code, the key idea of free software. The struggle would be thus transformed into a debate, in which different opinions might be held and a coexistence between opposed points of views could be achieved. The questions arising at this point are many: Can Internet be really free? Are there degrees of freedom in Internet? Are there other spheres in our social lives that are considered to be free? What is actually Internet? Has this debate any effect whatsoever on our non-virtual lives?
The following pages are just a collection of ideas and lessons, specially on the technical side, drawn from both my readings on the matter and my professional experience. I have started trying to find a clear definition of what is nowadays called “free software”. In the second section I move on to explain the types of licenses developed that allow a piece of software to be called “free”. The data available to the users depends on the type of license. Another issue to be kept in mind is data protection, a subject not developed in this paper but of great importance in the “freedom process of data”, since, it might be argued that there are some data that must be highly protected by Law and that should remain protected that way. Freedom and security are to very close terms. I have finally drafted a brief History of the free software in order to point out that software was originally free, and so was the Internet, when we understand the adjective, and attribute “free”, as synonym of “not subject to the control of a few and open to new contributions without fear of disturbing other’s rights in the process”. Again, data protection and privacy issues should be taken into account when speaking about a “free Internet”.
This is a very interesting subject, not only from the point of view of a jurist, but also from that of a citizen and member of this society. The meaning of fundamental freedoms and rights, the reaction of lawmakers and courts to an ever changing panorama, the speed of which requires fast but fair decisions, and the fight against the perceived control of a few companies over the rest of the world are just some issues I have come across during my incursions in this matter. I have tried to supply the readers and attendants with a big number, but not at all exhaustive, of links and bibliographical references along the text. Their conclusions, if any determining conclusion can be drawn regarding this “effervescent” field, might be very different from mine. But, isn’t this the most wonderful attribute of all debates?
A. What is „free software“? 
An attempt to define “free software”
According to the definition set by the Free Software Foundation (FSF), “free software” refers to the “user’s freedom to run, copy, distribute, study, change and improve software”. Therefore, the term free has nothing to do with price but with freedom, as it is clearly underlined by the FSF: “Think of free as in free speech, not as in free beer”.
Freedom of speech within algorithms?
It actually refers to four kinds of freedom the users of the software enjoy:
the freedom to run the program, for any purpose (freedom 0);
the freedom to study how the program works and adapt it to the users’ needs (freedom 1);
the freedom to redistribute copies, in order to help the community (freedom 2);
the freedom to improve the program, and release the improvements to the public, so that the whole community benefits (freedom 3).
A program is free software if users have all of these freedoms. But users are not in any way forced to introduce changes or redistribute the copy they have got. Thus, these freedoms are not duties.
A necessary condition for the freedoms 1 and 3 to be enjoyed is the accessibility of the source code, key to understand its functionality, to introduce changes and to publish improvements.
Before we continue I shall briefly explained what is source code and its counterpart, binary code. In many computer languages, programs are first written in a human-readable form that expresses the program logic. An automatic computer program translates that code into binary code (that is, ones and ceros) that is machine-readable. This program is called “compiler” and the process is called “compilation”. After the code is compiled, the binary form can be run in a computer. Many of the commercial programs that we use are distributed only in binary form.
Having the source code is essential for developers, since it is extremely difficult and expensive to reverse engineer the program to read its logic and underlying algorithms. Thus, source code is the key to the software’s maintenance as well as to its enhancement or improvement .
And what do these freedoms mean in practical terms?
Any kind of person or organization might use the program in any kind of computer system.
The freedom to make modifications means that there is neither need for permission or notification, nor for payment.
Lastly, the freedom to redistribute copies means this to be for free or charging a fee not fixed beforehand. In many cases, the price charged when redistributing is not related in any way to the price paid when the copy was purchased. This freedom also implies the availability of executable forms  of the program and source code, for the modified and unmodified versions. When the executable form of the program has not yet been developed or is simply impossible to be developed, a program might still qualify as free software. Furthermore, in case such a version is developed, the developer will be free to redistribute it under the same requirements we have seen up to now.
What are the features of these freedoms?
Irrevocable: no restrictions shall be added to the license;
Independent from way of distribution: everyone has the freedom to copy and change the software, even to sell it, regardless how she got the copies.
In order to protect these freedoms, some rules about how it might be distributed are acceptable, for instance, the rule of copyleft, that is, and in the words of the FSF, “the rule that when distributing the program, you cannot add restrictions to deny other people central freedoms”. Nonetheless, there is also free software that is not copylefted.
Finally, and in case you might have arrived to the opposite idea by now, free software may also be commercial. Why not? As long as these central freedoms are respected, the software is free. A different question is whether the big software concerns are happy with these rules, a matter that we will not cover here. Some examples of commercial development of free software is to be found in IBM  (Jikes and JFS), Apple (Quick Time Streaming Server) or Sun Microsystems (Star Office), amongst the traditional companies, and RedHat, VA Linux, Alcôve in France or ID-PRO, in Germany, amongst the new ones.
And, what is not free software?
The differences between free software and the following terms lie, mainly, on what is the user allowed to do with a piece of software. In a continuum, free software and proprietary software would be the two extremes, being all the other categories in between, nearer or farer to one or the other extreme according to the freedoms, understood as we have explained in point one of this section, they provide for (if any) and, thus, the availability of the source code.
To start with the easiest, it is plain  that proprietary software is not free software. Its use or redistribution is either prohibited or somehow restricted. Examples of proprietary software are all around us; we are just using some of it as we write this speech: Microsoft Word or Power Point are just two names among many others. Under the proprietary model, software vendor’s source code is kept as a closely guarded trade secret. The vendor has, thus, sole control of the contents of its software products as well as on the development of new features and maintenance.
Open source software:
Many of you are, surely, more familiar to this expression that to the “free software” one, but how many of you wanted to refer really to what “free software” means by using “open source software”? The differences between them made it important for the Open Source Initiative to issue a definition of “open source software” .
This includes free software as well as semi-free programs (i.e. Xv) and, even proprietary programs (Qt original license).
That is software that is not copyrighted. This does not imply its source code to be available. Even if its source code is available, this software will qualify as a special case of non-copylefted software, because some copies or modified versions of it might not be at all free.
As the Free software Foundation explains it, semi-free software is software that is not free, but comes with permission for individuals to use, copy, distribute, and modify (including distribution of modified versions) for non-profit purposes. PGP is an example of a semi-free program.
Generally used for packages the distribution of which is permitted but neither is its modification possible, nor is its source code available.
Therefore, free software is not a new technology, but an increasing popular licensing and distribution method of software. The software users may or may not take advantage of the availability of the source code, but it is certainly this availability together and the features of the licenses the most important determiners of how the product can be distributed and exploited, and what makes free software different in many ways.
B. Free data
In order to find out whether this ideas lead to the actual and unlimited availability of data, we will look closer at the concept of copyleft and the different free software licenses, including those developed beyond the FSF. Licenses are a very important matter for this kind of software, since it involves a compromise between several goals, among them :
To ensure the basic freedoms we have referred to (modification, copy, redistribution,).
To ensure some conditions imposed by the developers (citation of the author in derived work, for example).
To ensure that derived works are also free software.
The software developers will choose among the different available licenses in accordance to the degree with which they want to fulfill the mentioned goals.
Up to now we have talked of the freedom of redistribution, central to the definition of the free software. A program that might be developed by dozens of people without any relation between them but the goal to produce a better version of the code, how does it keep its nature of free throughout this process?
Some lines above we have mentioned the concept of copyleft. As we have already pointed out there, this is a general method for making a program free software and requiring all modified and extended versions of the program to be free software as well. Thus, it prevents anyone in the process from distributing the result as proprietary product. As we will see later, while explaining the different available licenses, copyleft is a “sticky” feature.
The code and the freedoms are legally bound through a two-step process:
First, the program is copyrighted;
Second, distribution terms are then added. These terms give the right to use, modify and redistribute the program’s code or (what is of the highest importance) any program derived from it (as long as the distribution terms remain unchanged).
1.2 Free software licenses
At the end of the nineties some organizations appeared in the free software community. These are carrying out an important task within the free software license landscape, by means of the definition of what a software license should have in order to qualify as a free software license. Most notably is the Debian Project .
The differences between licenses lie in the importance the author gives to the following issues:
protection of openness;
protection of moral rights;
protection of some proprietary rights;
compatibility with proprietary licenses;
compatibility with other open source licenses.
They all share the same requirement to keep the copyright notices attached to the software, including the own license, together wit a “No warranty” statement, or a waiver of responsibility of the copyright holders or third parties involved in case the program proves defective or causes any damages.
2. Free software licenses
Some of the most frequent free software licenses are:
BSD License (Berkeley Software Distribution) .
It mainly covers the software released by the BSD, but not only (as discussed below). Despite the appearance of permissive this license has, the Debian Project regards it as an example of free license . Why permissive?
In comparison with the previously mentioned licenses, the BSD one imposes almost no conditions on what the user might do with the software. It even does not provide for the obligation to include source code (!) and allows the redistributors to use the software for proprietary products. The authors only ask for their work to be recognized, thus, ensuring some free marketing (as the E.U. Working Group sees it). Therefore, if you like, you can bring BSD-licensed programs and code into the proprietary model, or, if you are the licensee of the BSD-licensed program, you can modify the source code and keep the resulting modified code completely private. This great freedom of use is very appealing for developers, but this ability to privatize is a bit at odds with the ideological goal of free software licensing to keep source code “free” from commercial control.
Open source licenses within this category, aside the BSD License itself (first used for the BSD Unix operating system form the University of California at Berkeley), are the MIT License, the W3C Software License (from the World Wide Web Consortium) and the Apache Software License (used for Apache web server).
GPL (GNU General Public License) and GLPL (GNU Library General Public License) .
Both license were developed by the GNU Project for its programs, but are to be found nowadays in software unrelated to this project.
In some ways, the GPL grants the licensee a great deal of freedoms, such as unlimited rights to use the software, to make and distribute copies and make derivatives. However, the GPL, loyal to its anti-commercial origins, has also added restrictions that are designed to keep the software from coming under any particular licensee’s control.
Its terms provide for:
The protection of the four freedoms referred to in the first lines of this speech: use, modification, redistribution and improvement of the program .
The availability of source code to each licensee of distributed binary copies; fees might be or not charge in the redistribution.
As we have mentioned before, the GPL has a “sticky nature” feared by many. The GPL says that it applies to “any work that you distribute or publish, that in whole or in part contains or is derived form the Program (i.e. the licensed program) or any part thereof...”. That means the following: if you take a few lines of GPL-licensed code and combine them with a million lines of your own proprietary code and then distribute copies of the resulting product, you have converted your millions of line of proprietary code into an open source product, subject to all the requirements of the GPL: source code access rights and unlimited royalty-free distribution rights to every licensee. In few words, an alarming result to companies developing software on an monopolistic fashion.
The GPL applies to the entire program, the source code of which this license is attached to. On the other hand, the GLPL only applies to the library protected, but not to the entire code, which might be, even, proprietary.
The widely known Linux kernel  is covered by the GPL.
MPL (The Mozilla Public License) 
As a reaction to the “sticky” nature of the GPL , there are Copyleft licenses that are less restrictive and clearer in their provisions. The Mozilla Public License, developed by Netscape, Inc. and which applies to the Mozilla web browser, is an example of this group.
The MPL is longer and ore detailed that the GPL and allows:
Release of code subject to a patent that the licensor does not control (although the licensor is required to provide notice of known patent issues).
Combination of MPL code with proprietary works without affecting the licensing of the larger work.
X Window System (X Consortium) license .
Very similar to the GNU GPL license but applicable to the software released by the X Consortium group, mainly oriented towards graphical user interfaces, or desktop utilities.
Almost-Open Source Licenses.
These licenses are considered to be between free source and proprietary licenses, since they allow access to source code, but do not grant the licensee freedom to make and distribute derivatives.
Examples of this category are Microsoft’s “Shared Source” program, under which academics may experiment with certain Microsoft source code, but are not allowed to distribute the results. Another is the Sun Community Source License (SCSL), introduced by this company for Java. The SCSL does not allow research use of Java unless the user pays license fees. In addition, it requires that any bug fixes and distributed modifications to its code have to be provided to Sun and subjected to Sun’s Java compatibility test before any commercial release.
Some vendors use a business model in which the same product is available under a “dual license”: The vendor available his software both under free software and commercial licenses. Thus, developer can choose between either one. For some companies this method is pretty convenient, since they do not want an obligation to make source code available. The strategy also benefits the vendor: The free software licensed copies spread throughout the user community knowledge of the product and its effectiveness grows.
These licensing method has so far worked best for products that are components of other products, such as MySQL  or Sleepycat Software.
In conclusion, freedom in Internet, if such exists, should start or starts with a freedom of choice between the different licenses available, a choice that depends on what the developer wants to achieve. One idea should be kept in mind when dealing with the issue of free data. Given the developing process of free software, as a result of contributions of many individuals, there is no way to be sure that each contributor had the rights under copyright law to supply such code. Collaboratively developed free software might include code added without permission and in violation of the copyright holder’s rights. Here lies, hence, the risk the free software movement might find on its way to make the source code available to everyone and promote a more distributed knowledge of the program development and, as a result, a more free Internet.
Before closing this section I will like to draw the attention to a very interesting web site:
http://www.wikipedia.org/wiki/computer-secience. This powerful online free encyclopedia might be very helpful to understand some technical terms I have poured on these pages.
Finally, after this extensive, but necessary, journey through the free software licenses, we will just give some important historical milestones in the development of the free software. And at the beginning was the Net completely free of property rights......
C. Free Internet: a still evolving issue
Historical development of free software 
Though the phenomenon of free software might seem to be quite recent, it is actually how it all, or almost all, began in the world of software development. In fact, in the decades of the sixties and seventies, software and source code were freely distributed not only within academic circles but as well in forums like IBM SHARE or users groups such as DEC DECUS. Proprietary software, despite its late development, rapidly evolved throwing free software almost “out of the road” until recently, when the industry is reconsidering the advantages of free software.
Here are some dates to better understand this process:
1950’s and 1960’s: Software is distributed together with source code and without restrictions.
1960 (April): RFC number 1 describing the first software for the Internet (then ARPANET) is published. Its free availability, in particular of its protocol specifications, was a key factor to the later development of Internet.
1972: Source code is freely distributed in academic circles. Examples: the MIT (Massachusetts Institute of Technology) distributes HACKMEN (an assembly language).
1978: Professor Donald E. Knot from the Stanford University begins to work on TeX, a typesetting system distributed as free software.
1983: Richard Stallman writes the GNU Manifesto, in which he calls for a return to the public sharing of software and source code.
1984: The GNU Project begins. The goal is to build a complete free operating system, taking UNIX as a model.
1985: the MIT based X Consortium distributes the X Window System as free software covered by one of the less restrictive free software licenses.
1985: The Free Software Foundation  is created.
1989: Cygnus, the first commercial company devoted to provide commercial support for GNU software, and open source software, is funded.
1990: The FSF announces it intent to build an Unix-like kernel, called GNU Hurd, in order to fill in the last major hole in the GNU series of software, thus, creating a fully free development system.
1991: The Jolitzs write a series on BSD. As a result of the open source software developed and collected at the Centers for Software Research of the University of California at Berkeley, it was possible to have a complete (and free) BSD operating system. It is also the beginning of the BSD family of free and open source operating systems (NetBSD, FreeBSD, OpenBSD).
1991 (Aug): The Finnish graduate student Linus Torvalds announces that he has been working on an open source Unix-like kernel, using GNU tools.
1991 (Oct): Linus Torvalds publicly releases the source code for his Unix-like kernel, named Linux.
1991 (Dec): Linus Torvalds announces the first self-supporting release of Linux, on which developers can work, without using any proprietary tools or operating systems .
1992: The U.S. Air Force awards the New York University (NYU) a contract to build a free compiler. The NYU team chooses the tool GNU GCC for code generation. The product is named GNAT, which stands for GNU NYU Ada Translator.
1993 (Aug): Ian Murdock created a new Linux-based distribution called Debian  GNU/Linux, developed by a group of volunteers around the world.
1994: Ada Core Technologies is incorporated. Its core business is the development of GNAT  and the selling of service support.
1994: Marc Ewing begins the Red Hat GNU/Linux distribution .
1995 (Apr): First official release of Apache  is distributed. Apache is the web server 58% of the web sites use nowadays, making it the leader of its commercial niche, well ahead other proprietary solutions.
1996: First Conference on Freely Redistributable Software, Cambridge, Massachusetts, U.S.A.
1996 (Oct): The KDE  project is launched in order to address usability problems of free software, mainly by developing desktop utilities.
1997 (Jun): Eric S. Raymond presents his paper „The Cathedral and the Bazaar“  on why Linux software development model works.
1997 (Aug): GNOME , another desktop environment tool project, is born.
1998 (Jan): Netscape declares its intent to release the source code for its Navigator browser. It is finally released in April the same year .
1998 (Feb): The term “open source” is coined by Chris Peterson and registered, in order to act as a trademark for free software products .
1998 (Jul): more than 300 volunteer developers working on more than 1,500 packages release the new version of Debian.
1998 (Aug): Linus Torvalds and Linux appear on the front cover of Forbes Magazine.
1998 (Oct): IBM decides to test open source by using Apache.
1998 (Nov): “Halloween” documents (attributed to Microsoft) are leaked to the public . The documents analyze strengths and weakness of free software and Linux. The FUD (Fear, Unknown and Doubt) technique starts to be developed by big software concerns.
1999: New and more secure versions of free software are developed around the world.
1999 (Oct): The French Senate publishes in its web site a proposed bill on the preferable use of free software in the Administration , the result of an open forum in Internet. The Lafitte, Trégouet and Cabanel project (known after the name of the senators who proposed it) has not been adopted, but signals a starting point in the E.U. Governments movement towards free software environments.
1999 (Nov): The German Government declares its intention to fund a free software development project. The 163,000 Euro-fund was awarded to German UNIX User Group , the goal of which was to improve the software GNU Privacy Guard, a privacy application.
1999 (Dic): The Congressman Walter Pinheiro in Brazil presents a bill on Free Software .
2000 (Mar): The E.U. Working Group on Libre Software  presents its paper in front of the European Commission.
2000 (Apr): the French Congressmen Le Déaut, Paul and Cohen propose another bill on free software , similar to the above referred, proposed by Lafitte, Trégouet and Cabanel.
2001 (Jan): Amazon.com switches to Linux .
2003: Linux WorldExpo records a high attendance: private users as well as business, the presence of which grows every year .
Historical development of Internet
As mentioned above, Arpanet is the direct ancestor of Internet. this project was created by ARPA, the Advanced Research Studies Agency of the Defense Department of the United States. Both its design and the location of its original nodes, universities, were based on the idea of an open network.
Arpanet was developed in order to help the communication between different university software research teams, sponsored by the Defense Department, and allowed them the shared use of powerful computing devices only available in some of these universities. The increasing capacity and speed of the computers threaten the further development of Arpanet thought to have been outshined. Nevertheless, one of these research groups found out that Arpanet could help them carry out their studies and started to develop a system that allowed computers to communicate with each other: the TCP/IP protocols by Cerf, Kahn and Postel.
Therefore, the first Internet developers, all coming from the world of university academics, tried to establish a free network that would allow them to communicate their results with other colleagues. And here lies the reason why Arpanet, although sponsored by the Defense Department, “born free” of any control, only driven by the aim of its continuous improvement.
During the seventies and eighties, the Internet spread beyond the academic circles involving other social groups and cultural movements. Every new group contributed with new applications that were published for general notice and common development. This free exchange of knowledge coexists nowadays with a more commercial use of the network. Does this mean that Internet is no longer free? On the one hand, the commercial use of Internet has opened the network to social groups that might have never been aware of this tool otherwise . Technology and news can be more easily spread through the Internet and reach a wider number of people around the World. The benefit is two sided: the Internet gains in variety and the society, in general, enjoys from an opened window towards issues and ideas once concealed to a limited group of wise men. On the other hand, Internet has developed into a worldwide agora and software has become a valuable good. The early practice of sharing ideas and results between developers in order to get better computing tools does not match any longer with the notion of property and benefit. Property sets limits to the rights of third parties . Some algorithms are subject of copyright and may not be freely distributed without the previous agreement of the holder of the rights, who might have grounded his power on that specific group of formulas. And although the need to improve the existing software and network is felt on both sides, the way to achieve and the finality of this improvement differs in either side. Nevertheless, this tension is a creative one, since it helps both sides to keep on looking for better solutions to similar problems. An important task now is to keep a close look on the regulations passed by Governments that could hinder the development of Internet, that could try to restrict the flow of creativity attached to the flow of information available.
 For the drafting of this section we are mainly using material from the two very illustrating articles: “The Free Software Definition”, written by Richard Stallman. Available at the FSF web site, http://www.gnu.org/fsf. “Free software/Open source: Information Society Opportunities for Europe?“ written by the Working Group on Libre Software, created at the initiative of the Information Society Directorate General of the European Commission. Available at http://eu.connect.it/paper.pdf, the latest version available of which was issued in 2000.
 I sincerely thank Mr. K. Landy for explaining to me this technicality in such an elegant fashion.
 Binary forms.
 The Sun Community Source License (SCSL) provides an interesting example of the development of proprietary software within a free software or open source software model, thus, blurring the lines between these two terms. See FAQ on the SCSL at: http://www.sun.com/software/communitysource/faq.html.
 Open Source Initiative official definition at: http://www.opensource.org/osd.html. The Free Software Foundation also tries to draw the line between these two expressions at: http://gnu.org/philosophy/free-software-for-freedom.html
 Taken from the paper “Free software/Open source: Information Society Opportunities for Europe?“ which refers to Peter L. Deutsch: Licenses for freely redistributable software, in Proceedings of the First Conference on Freely Redistributable Software, Cambridge, Massachusetts, U.S.A., 1996.
 http://www.debian.org/social -contract. The Open Source Initiative has also defined “open source software” licenses, but these do not qualify as free software according to the definition we are using here. In any case, the work this group is developing is available at: http://www.opensource.org.
 Available at http://www.frebsd.org.
 See note 7. Social Contract point number 10.
 Available at http://www.gnu.org/
 The GPL-license clearly provides for the following: “Each time you redistribute the Program (or any work based on the program), the recipient automatically receives a license from the original licensor to copy, distribute or modify the program subject to these terms and conditions”. That is, as a GPL licensee, you may distribute the code or its derivatives only under the GPL. Unlike code under a BSD-Type license, a GPL-licensee cannot bring code into the proprietary model. Rather the GPL is designed to lock each licensee into the GPL model for any further distribution forever. As copyright is a right to exclude others, the “Copyleft” is a requirement that licensees be included in the development, distribution and source code access rights.
 The kernel is the part of the system that allocates the machine’s resources to the other programs the computer user runs.
 Some even called the GPL a “viral” license or a kind of IP cancer.
 Available at http://www.xfree86.org and at http://projects.openresources.com/libresoft-note/libresoft-note-en/node108.html.
 We encourage once more the reading of the highly educative paper produced by the E.U. Working Group on Free Software: http://eu.conecta.it/paper.pdf. This section is based on this paper.
 More information on Linux at: http://www.linux.com and http://www.kernelnotes.org.
 http://tuxedo.org/ (search for the section Writings and the name of the article: cathedral-bazaar)
 More information at: http://www.mozilla.org.
 An interesting discussion on this term might be read at http://www.gnu.org/philosophy/free-software-for-freedom.html, referred to in the previous section.
 More information in Wired News at: http://www.wired.com.
 Highly interesting are the following sites: International Telecommunication Union, World Telecommunication Development: http://www.itu.int/ Observatory on the Information Society: http://www.unesco.org/webworld/observatory/index.shtml Human Development Report 2001: Making New Technologies work for human development: http://www.undp.org/hdr2001 Association e-Développement (AEDEV): http://www.aedev.org