In trying to think about the success factors for Open Source (OS) projects, and evaluate their character and structure, as well as thinking about extending this idea to other areas, I had the insight that the essential character of OS project organization is anarchy. As a political/intellectual movement, Anarchy is probably the most pure form of Libertarianism. Forget any associations you may have with the idea of creating anarchy in communities or societies by throwing bombs and other disruptive acts, since these are both factually incorrect, and have nothing to do with what Anarchy advocates. The correct association is of anarchy with "a state of nature", the Garden of Eden, if you will.

Humans, being highly social animals with highly advanced systems for communication of symbolic knowledge, have the ability to impose rules of all sorts on this original state. In principle, there is nothing wrong with this, but history shows many examples where "the rules" become highly oppressive. In tribal societies, the social unit is a small group where social "norms" can operate effectively, and it can be argued that the "norms" are essential for the survival of the tribe, but human development did not stop there. With the development of agriculture, the stage was set for creating hierarchical structures, monetary systems and large scale warfare (i.e. beyond inter-tribal conflicts for territory).

It is well know that Libertarian thought is pervasive in the highly technical software development community, and it is easy to see the attraction of these ideas to a class of highly intelligent, somewhat individualistic people. Add youth to that, and you get a lot of contempt for conventional systems of power and authority. In the beginnings of the software industry, there wasn't much of a market for additional copies of specific programs, and a lot of development happened in academic and other research labs, so there wasn't much thought or attention from the capitalists. Programmers freely shared their code with anyone who asked, and nobody thought about cashing in by selling millions of copies of a program. Richard Stallman created the GPL in reaction to the way code sharing was being closed down by the potential to cash in by selling code over and over.

Now that the OS concept is getting mature and people are starting to do academic studies of the organizational structures developing around OS, and others are attempting to extend this model to other forms of intellectual property (IP), they are finding a conspicuous lack of planning or formal thought in these areas. Someone starts with an idea and writes some code, then shares it and begins to develop a small tribe of followers, and delegates some of the core responsibilities within that tribe. Within these groups, social norms control the interaction, and sometimes things go badly, but because the code itself is in the "commons", anyone can "fork" their own project at any time. Of course, this doesn't happen often because for the fork to be viable, it must represent a real split in the community, not just one or a few upset people. Typically, the original developer (or small team) retains control of the project for as long as they care to, but this is just another social norm that acknowledges the contribution, skill and insight that it took to get things started. It is often stated that OS project leaders are "dictators", but they must be benevolent because good will is the only thing that holds the tribe together (ok, so there is also the large amount of code to maintain if you do start a fork, but there are many OS projects to work on these days).

All of this is the essence of an anarchistic organizational system. Yes, formal structures are developed and put in place, but only with the tacit support of the community. It only works because everyone is free to participate or not, according to their desires and interests. There would be no debate about any of this if we weren't embedded in a system of market capitalism where value is equated with money, and money is necessary for each of us to be able to live and make choices. Some of the most vocal critics of the GPL actually try to claim that OS projects take away their livelihoods by making it hard to make money writing software. I won't go into all the ways this is wrong here, but it does open up the question of how OS developers can support themselves and their families.

Clearly, the large and growing community of OS users, both individual and corporate, demonstrate that they value the contributions made by core developers, and they often find ways to support them. I take the position that both in the interest of fairness, and to better promote the wide and open sharing of IP of all types, that we need to have more and better ways to compensate significant contributions, without having this dominate. Yes, there are many other motivations for people working on IP projects, and many can work on OS projects with the approval of employers who need features and functions that might not otherwise be developed, but many skilled people cannot afford to forgo employment to work on OS full time, or find sufficient time to do this as a second job. Academics often need publication credits to satisfy their institutions and get grants, but this also tends to distort the work to satisfy the sponsors.

The bottom line is that while monetary systems and markets work well to efficiently distribute scarce commodities, they also tend to simplify complex systems of values into a single dimension, and they are particularly bad at promoting the efficient development of IP resources that gain their greatest value the more widely they are shared. It should be clear to most of us by now that this one-dimensional value system becomes non-functional in an information economy, as well as undervaluing the diversity and quality of the natural environment necessary for our long-term survival. The way forward will involve the emergence of new value systems based on sharing of information. To get there from here, we need to operate in the context of market capitalism, and actually exploit it to fund the transformation. This will involve convincing those who control the money to fund the rapid development of the IP Commons for the benefit of everyone.

I was prompted to write this after reading a paper I found at an on-line journal, First Monday:

The Institutional Design of Open Source Programming: Implications for Addressing Complex Public Policy and Management Problems by Charles M. Schweik and Andrei Semenov

In their conclusions, they talk about attempting to extend the OS development model to other areas, which I fully support. On the other hand, I also think there is something about software that distinguishes it from other types of intellectual property. Specifically, it is the only type of digital information that is both descriptive and functional. Papers, music, videos, and even databases are not fundamentally "engineered" designs where the proof is largely in the way it functions. This places certain constraints on the process and facilitates "objective" feedback from the community. You don't need to look at the source to know if it works or not, just to run it. This is true in a less direct way with the design and engineering of hardware and other physical devices. My instinct tells me that most types of IP will be harder to establish in OS models, but no less valuable when widely shared. This will make it even more critical that these projects are carefully evaluated for their quality and effectiveness, or most of them will never reach the critical mass necessary to be widely useful.