Steven Weber’s excellent book, The Success of Open Source is a superb complement to Yochai Benkler’s classic essay, Coase’s Penguin. Benkler looks at peer production as an economic system and concludes that it has become a third major form of organizing production, alongside the market and the firm. Weber takes a closer look inside the open source production process, and provides a fascinating analysis of how and why it works:
- the origin of open source software
- why people participate
- how projects are organized
- how open source fits into surrounding organizational and economic structures.
By doing this, Weber reaches a variety of interesting observations and conclusions:
- Counter to the myth, the open source development process is not a teeming bazaar, with “bottom-up” self-organization composed of local signals. The largest and most successful open source projects have identifiable, hierarchical organizational structures, with a leader and/or inner circle, up to a few hundred active contributors, and a much larger group of occasional participants.
- While Open source licenses protect the right to “fork”, to take the codebase off in a different direction than the original project, projects stay together more than a skeptic might think. Weber observes that project leaders depend on developers and developers depend on the community. The ability to get more done together than separately.
- Developers contribute to open source projects even though most users are “free riders” who benefit from the software, and contribute little or no code. This is less of a paradox than it might seem, since software is a “network good” that gains value the more people who use it. The more people who use a program, the tbugs that are reported and fixed, and the more robust the system becomes.
- Since the origins of the phenomenon, there have been different approaches to licenses. The West Coast, Berkeley-style licenses are easy-going about the ability to include open source software in other, non-open source code, so long as credit is preserved. The East Coast, Free Software Foundation GPL (Gnu Public License) is strict about requiring that redistributed code must always be free software, and any software including free software must be distributed by GPL
Perhaps the most insightful conclusion Weber draws is the relation between open source and intellectual property. Weber observes that open source redefines property around the right to distribute, not the right to exclude.
Weber is able to make this observation because he avoids polemic. Weber doesn’t try argue that open source software is good because intellectual property is bad. And he doesn’t argue that that open source software is bad because intellectual property is good. Instead, he is able to observe how open source redefines property itself.
Weber’s pragmatic analysis leads him to focus on the vibrant intersection between open source production and traditional business, with a look at a variety of hybrid business models, from IBM’s focus on hardware and service, to Red Hat’s packaging and branding, to MySQL’s service and customization, to Apple’s addition of proprietary chrome and polish. Weber predicts continued evolution and innovation and this boundary.
The book was published in 2004, and so it misses one of the most interesting trends in the last couple of years — the rise of open source software that’s not just for hackers. Netscape/Mozilla is included in the book as an example of failure. Weber looks around at Linux, Gnome, KDE, etc, and concludes that open source software may never be able to make software that works for non-hackers. This was before before the breakout success of Firefox, and the popularity of GAIM, an instant messaging client with a consumer-quality interface.
Weber examines the brash and blunt hacker culture, with its focus on technical decision making through vehement debates on project mailing lists that hash out solutions to technical problems and decisions about technnical direction. I wonder about how the culture will evolve as interactions grow with non-geek users, and hybrid companies face decisions that have external constraints driven by customers.
Towards the end of the book, Weber speculates about how the organizing methods of open source software might affect the production of other kinds of goods — writing, music, biotech, business ideas. I thought it was interesting, but less substantial than the parts of the book focused on open source itself, with analysis based on observation.