In 2000, IBM announced their intention to open source their AFS product. A number of us, me among them, got together to try to set up some organization to be ready to receive the code drop and run with it. At the earliest gathering of these folks, I got up in front of group, some of whom I'd just met, stood in front of a whiteboard, and sketched out a plan for an organization I thought would work. By and large, that organization is the one that got built.
Fast-forward a few years. One flaw in my plan became apparent: I wrote something which vested membership in the guiding board in people, not their organizations. But those people were not always motivated to get things done, or to step aside when they failed to. We met again. I laid out my complaints, offered my resignation, and left the room. It was not accepted, but the shakeup yielded a reconstituted board, and we again moved forward, if slowly.
Over time, another flaw became apparent. Another group of people, of which I am also one, vetted code changes and did all the release engineering work to get things shipped. Except we weren't all on the same page. I assumed the responsibilities of this would be implicit. I was wrong.
Finally, lacking any formal resources, things got done when someone volunteered and had time, or paid and contributed the work.
A draft plan, basically the second proposed structure for an organization which would exist as a vessel to steward OpenAFS, has been available for comment for a few weeks. Like the previous plan, this one also came from me, albeit informed by a new, different, and hopefully better set of sources and experiences. Among them, social media and open source communities have been talking about many of the activities needed to build a base for such an organization, and I've been privileged to be part of Ohio LinuxFest '08, Podcamp Pittsburgh 3, and the Google Summer of Code Mentor Summit over the past weeks.
One can only hope finally we get it right.