Open Source, Open Society?
Felix Stalder,
A talk given at an Open Soure event in Toronto, November 1999

 

As far as I can see, there are two characteristics of the development process of "open source" software [1] that are likely to influence the resulting products in ways that go beyond strictly technical questions: the heterogeneous developer base, and the particular licensing structure. Open source software is the result of a collaborative effort of different people who each pursue diverging personal and collective agendas when participating in this process.

By "agenda" I mean simply someone's reason(s) to do a certain thing. Some of the reasons to engage in open source development are peer recognition, efficiency, aesthetic pleasure, financial gain or a particular social/political belief. Proprietary software is also developed by a number of different people, who arguably work on it for many different personal reasons (being paid is but one of them). However, there is--and this is the difference to open source process--a single dominant collective agenda: the agenda of the company that owns the software and hires the programmers. For a publicly traded company, this agenda has to be to maximize value for its shareholders. At the end of the day, this single collective agenda overrides all others.

The combination of a single agenda that lies outside of the software itself and opaque source code makes it easy to put features into the software that are controversial, or even unpopular, but serve the agenda which dominates the developmental process. If Microsoft (or Sun, or Oracle, or Apple, or...) reaches the conclusion that its interests are best served through by entering into a secret partnership with, say, the NSA (US National Security Agency) then the terms of this partnership will be implemented by the programmers, no matter if they personally belief this to be a good thing or not. Examples of controversial, hidden features are abound: back doors in encryption software, such as the controversial "NSA key" recently discovered in Microsoft NT stations, or the audio software "realplayer" which sends data about the user back to the software company, real.com. Both features reflect overarching agendas of the developers which are unchecked, and cannot be checked, by other developers or users. Open source software is very unlikely to contain such hidden features. Not only because it is open, hence the features would be visible to literate users, but also because the agendas of the people working on the development of the software are very diverse. Even more important is that in the open source development there is no mechanism by which someone could force someone else to adopt something against his or her own personal conviction, no matter what this convictions is. Given the impossibility of imposing an overarching agenda it is unlikely that there will be features embedded in the code that clearly promote any particular non-technical goal, such as gathering data for marketing purposes, or improving relations to government agencies. Open source software represents an original model of ownership. This model is based on the GNU license agreement developed by Richard Stallman [2].

This license mandates "that anyone who redistributes the software, with or without changes, must pass along the freedom to further copy and change it." [3] Effectively, this guarantees that once a piece of software is protected by this license, its current code and its later versions cannot be taken out of the public domain anymore. The traditional model of software licensing we all know well. The developer licenses the software to the user under conditions which usually include that the user cannot distribute or modify it, nor use any pieces of the source code for other products. Hence the code can never be put into the public domain. The result of this open source license is not only that many different people can work on the software for many different reasons, but also that the software becomes much cheaper because its impossible to produce an artificial scarcity. With the Internet as distribution mechanism this software tends to become gratis because one single freely available copy is infinitely reproducible at basically no costs. These two characteristics of "open source" software development process tends to result in software that is "cleaner" and cheaper than proprietary software.

 

Does this matter? It does. Software needs to be clean. Computers and software can be thought of as amplifiers. They amplify the user's agenda by giving her access to means of, say, communication that she would not have otherwise. But computers and software also amplify the agendas of their makers. For example, the realplayer allows 16 million users to listen to whatever they personally find worth listening to, the software amplifies their power to gain access to recorded sounds that are stored on-line. On the other hand, all these 16 million players also promote the agenda of their developer, real.com, which now has 16 million 'agents' in the field reporting back in the users listening habits. Effectively, the realplayer amplifies 16 million user agendas once, and one company agenda 16 million times. Hence it empowers each user a little bit and the developer tremendously. The same can be said of the Windows operating system. Open source software reduces this imbalance. The various agendas of the developers cancel out one another as they meet on a relatively restricted common ground: the development of technically superior software. Consequently, open source software empowers the user vis--vis the developer for the simple reason that the nontechnical motivations of each individual developer become less important because they are checked by others who can not be assumed to share these motivations. Checked from a wide ranges of angles, the software becomes not only more stable, but also more clean or neutral.

Paradoxically, this political neutrality is a radical political feature in a context where software that is biased towards the developer is the normality. Software needs to be cheap. While clean software addresses the imbalance of amplifying power between the developer and the users, cheap software allows more social groups to use that power than simply those with money. At the centers of technological development this is not such an important point because the connection between knowledge and money is more direct. The situation is different in developing countries where knowledge is more abundant than money. Open source software, because it is much cheaper, allows more people to use the amplifying power computers. The decision of the Mexican government to use Linux and not Microsoft in its schools was, at least partly, motivated by the fact that the lower software costs made it possible to install more machines. For the time being, the low costs which increase its accessibility is offset by the still very high knowledge necessary to make use of the much of the software. However, it might be a temporary issue as we can expect the software to become more "user-friendly" and the required knowledge to become more distributed. The more ubiquitous computing becomes, the more important is it that the software is clean, that is, free of unchecked special interests. The best way to achieve this is to make very diverse interests have access to the same code. At the same time, the more essential computing becomes for the conduct of everyday life, the more is it important to widen the access to the basic tools. Making the software freely available and opening up its code for inspection and change transforms the character of software from a commodity into something more like an environmental resource of the Internet similar to air in the physical environment. Everyone has access to it and everyone is allowed to check its contents. Such a transformation is in itself positive as it helps to reduce the imbalances of power between the developer and the user, and between the rich and the comparatively poor.

However, what the effects of this leveling of the playing field will be on other areas of society is more ambiguous. What seems likely is that it contributes to accelerate the much more general shift from a commodity to a service economy. Those who focus on services can do very well, even if they do not own the software which they service, as the case of Red Hat, Inc. indicates. In a limited sense, open source code is like the legal code. The code is openly published and accessible to everyone. Nevertheless due to its complexity, most people do need to rely on a professional who can interpret the general rules in the light of their own unique situation. What seems unlikely, though, that open source software represents in itself a new production paradigm--a "gift economy"--which can transform the fundamentally capitalist character of the (new) economy [4].

[1] I use the term open source software of all types of software allows the user to modify and freely pass on its source code.
[2] http://www.gnu.org/copyleft/gpl.html
[3] http://www.gnu.org/copyleft/copyleft.html
[4] http://www.firstmonday.org/issues/issue3_12/barbrook/index.html