Thursday, January 25, 2007

Cause and Effect: Shared Governance Brings Community

In his blog today, Matt Asay makes an impassioned plea for the open source world to draw a sharp distinction between companies that are sincere about open source, and those who are not. He argues the firms that fully embrace open source will succeed — while those that are half-hearted or “treat open source as a ball-and-chain” will fail.

Based on my research on firm open source strategies over the past 7 years, I am sympathetic to the argument. My hunch is that in most cases there is a structural reason for insincerity: Microsoft (or another big proprietary software vendor) is insincere at open source and fails, but the root cause is that a full embrace of open source would cannibalize its business. Clay Christensen calls this The Innovator’s Dilemma.

However, I take violent exception to the throwaway line at the end:

Open source without community is a sham.
IMHO this is confusing the cause with the effect. There are 100,000+ SourceForge projects without community. Are they a sham, too? Or are they just otherwise sincere people that failed to attract an audience?

For the past 4 years I’ve been researching how firms work with to open source communities, working with Siobhán O’Mahony who did her Stanford Ph.D. on the subject and was the first academic sociologist to study the phenemonon. We’ve concluded that open source communities have three dimensions:
  1. intellectual property
  2. development practices and
  3. governance.
IP (i.e., OSI licenses) and development practices (e.g. CVS, Bugzilla etc.) fall into a fairly narrow range. However, the governance decisions (both transparency and accessibility) that firms make for sponsored open source projects strongly influence whether outsiders decide to join a community.

The obligation of a real open source company is to create governance structures which allow external participants to feel that they have the full right to participate and influence the outcome of the open source project; the best example of this is what IBM did with Eclipse. If no one shows up, then it might be because the sponsor failed to share control — or it might be because the technology has a small potential user base or because another project has already attracted the largest pool of potential contributors.

Matt is a friend, and one of the most insightful people I know on open source strategies, but I think he’s plain wrong here. It may have just been a minor slip of the fingers.


Matt Asay (Pronounced "Ay-see") said...

Not really a slip of the fingers, but certainly a slip of the rhetoric. :-)

I guess I'd classify all those single-person Sourceforge posts as community-driven. It's more commercial open source that I'm calling "sham" on (to the extent that such projects lack community). And I'm pointing the finger at myself as much as anyone else.

All that said, I appreciate you taking me to task, and all the work you've done that has informed my opinions on open source.

Joel West said...

With the clarification, I think we could agree that “where there’s smoke, there’s fire.”

If a project is doing a bad job at attracting an outside community, maybe there’s some innocent reason. But with a commercial sponsor involved, often (usually?) it’s because the sponsor won’t let go of enough control to attract outsiders.

Sometimes a perception of being faux is enough. Even if the project claims to be open, if contributors don’t believe it, they’ll stay away.

Still, in my Eclipse research, I got the sense that if the business stakes were high enough, outside firms would make the effort to push to get their issues addressed rather than give up.