You've seen the lists: "Here are the 7, or 12, or 34 things you must always do, or never do, to have a successful online community!" These lists are well-meaning, and the things they describe do matter, but they rest on a false premise. Community can't in fact be created with a recipe. As Exhibit A, StackOverflow, a burgeoning 'question-centric' community has required subtle adjustments in its communal structure almost from the beginning, adjustments that are never captured in simplistic lists.
For most software, the critical period is the last month before launch -- compile, check, fix, compile again. For social software like StackOverflow, the critical period is the first month after launch, when the first users join, get their bearings, and have their first big argument. (No argument == no community.) This is because the run-time environment for social software is the minds of its users, and the incentives and disincentives created for those users early on have a huge effect on eventual cultural norms. The only to get those norms right is for the people running the site to understand what's going on in the minds of the people using it. In StackOverflow's case, this meant that a site run for programmers has to be run by programmers, so that they can understand what motivates their users.
In this talk, Atwood and Shirky will discuss about the kinds of lessons about online community that *can't* be applied like a recipe, focusing on complex and messy stories from the real world, drawn from StackOverflow and elsewhere.
Jeff Atwood lives in Berkeley, CA with his wife, two cats, and a whole lot of computers. He has a particular interest in the human side of software development, as represented in his Coding Horror blog. Jeff is also CEO of Stack Overflow, a fledgling Q&A community for programmers, with his business partner Joel Spolsky.