Bazaar Model

From P2P Foundation
Jump to: navigation, search


Mathieu O'Neil:

"The idea of “Bazaar Model” was the first conceptualization to address a productive scheme which was exclusive of Linux at that moment. Eric Raymond, author of the metaphor in 1997, contrasted the notion of a cathedral -a hierarchical and closed architecture-with a noisy bazaar, where each visitor touches, tastes, puts in and puts out what he wants.

- Linux came to overturn much of what I thought I knew. I had been preaching the gospel of UNIX for years, the gospel of small tools, fast prototyping and evolutionary programming. But I also believed that there was a certain critical complexity which required a planned and centralized approach. I thought the larger software (operating systems and really large tools like Emacs) needed to be built like cathedrals, that is that it should be carefully built by geniuses or small bands of mages working locked and bolted, without releasing early beta versions. The development style of Linus Torvalds ("release early and often, delegate everything you can, be open to the point of promiscuity") surprised to me. It was not in any way a cathedral building. In contrast, the Linux community was more like a noisy bazaar of Babel, full of individuals with various aims and approaches (faithfully represented by the Linux file repositories, which can accept contributions from anyone), where there would emerge a stable and consistent system only after a set of widgets. (Raymond, 1997:1) (


Mathieu O'Neil:

Miquel Vidal, in his excellent work, systematizes the ideas of Raymond noting three features of the Bazaar model: 1) Release early and often 2) distribute responsibilities and tasks as much as possible 3) Be open to the maximum. In turn, he stresses that the model does not work "when projects begin from 0 or when involves small groups too heterogeneous or with disparate knowledge" (Raymond, 2000: 55-56). But the interesting thing is that he is the first who reflects on the relationship between this productive modality and capitalism. His notion of Noncommand Cooperation takes account of that.

In any case, with or without Bazaar, and beyond the demonstrated success of free software in organizational and technical matter, it defies the interested and commercial logic which seemed definitively installed on social issues. Someone might argue that the cooperative processes are not new in advanced capitalism, and that in fact are a critical part of post-Fordist model of organization. But the latter requires subjected cooperation, oriented only to the extraction of profit, in any case being self-determined. The novelty of free software is operating a model of non-command cooperation. There is no direct commercial interest; there is pure general intellect, unruly and free from command. (Vidal, 2000: 56)

In turn, Vidal wittily grasped the increase of the companies´ involvement in the collaborative production when it was just beginning.

Some large companies have started hiring hackers (which is not new) to develop free software (this is new). Works previously done without direct economic interest now begin to be funded by companies. Projects previously motivated by the hackers´ need or desire and the community of free software users, non-market interests, can now begin to be marked by the needs, rhythms and priorities of the companies which fund these projects. Little companies whose incomes were based on services related to free software have suddenly turned into large companies involved in the stock market with risk capital. Some companies based on free software business are buying smaller companies and in turn are bought by larger ones, resulting in the creation of great empires. That bustle of sale includes strategic sites for the community as a media or software repositories: Andover buys Slashdot and Freshmeat, VA Linux buys Andover, RedHat buys Cygnus, etc. (Vidal, 2000: 63)

But what were the implications of such advance? With a striking ability to intuit the future, Vidal reveals his concern.

So far, in the open source community this is not seen as a threat, even as a problem, but on the contrary: some people have worked hard to convince companies of the capitalist viability of the model, and now the fruits are beginning to being harvested. How can we now oppose to companies making money by the model, if they keep the rules of the game, that is, if they produce or finance free software? Neither we have perspective nor has enough time passed (just two years) to assess what is going to happen after the massive involvement of strong and transnational capital in free software. My personal opinion is that, unlike other issues in which approaches are critical and very cautious (such as patent law), in this crucial issue there is excessive faith in the goodness of the market and free trade. (...). Cooperation without command can easily switch to subjected cooperation, cooperation with command. (Vidal, 2000: 64)

However, despite this observation, Vidal believes there is an irreconcilable contradiction between the collaborative production and the business logic.

Moreover, the absence of command, of corporate or hierarchical control, seems a sine qua non condition: where reappears command | whether in the form of proprietary interest, whether in authoritarian form|, the model will fall, wither and finally disappear. (Vidal, 2000: 56)

As we will see, the reality has rebelled against this assertion. The model of non-command cooperation – of course, the term is no longer appropriate- , not only does not wither but flourishes, enthusiastic, in the most various forms of capitalism." (


From note 1 and 2 of a 2010 submission to the CSPP journal, "Collaborative Production', by Mathieu O'Neil.


Excerpts from Vidal:

  • VIDAL, Miquel (2004)[2000] Cooperación sin mando: una introducción al software libre en Gradin, Carlos (compilador) :(){ :|:& };: Internet, hackers y software libre, Editora Fantasma, Bs. As.

Excerpts from Raymond: