A few months ago, I spoke with Joel Spolsky, CEO of Stack Overflow Internet Services, about Stack Exchange, software that powers Math Overflow, a collaborative tool that allows math researchers to collectively solve complex equations online. Stack Exchange is best described as a tightly structured question and answer software, designed to conscientiously facilitate efficient problem solving. Spolsky and I had discussed the possibility of future iterations of Math Overflow's wildly successful model, and this week Stack Overflow launched a new Physics beta forum, covering questions ranging from "home experiments to derive the speed of light" to "why isn't dark matter just matter?"
I spoke with Robert Cartaino, community coordinator at Stack Overflow, to ask what other projects he had planned for the future. His answer was interesting: user engagement shapes what forum the Stack Overflow team will craft next. Each topical idea for a site -- be it physics, writing or cooking -- is entirely crowdsourced from the user base; Stack Overflow merely builds the infrastructure.
According to Cartaino, the Physics site started out as a proposal. If a group of people can show that an enthusiastic community is ready to use the site, Stack Overflow will create it.
Site proposals are vetted through Area 51, the Stack Exchange Network's staging zone, where users "help sites get off the ground by defining what's on- and off-topic, recruiting ... experts, and committing to the site's success." Site staging follows a fairly specific process: users "follow" the site until it has a critical mass of users, narrow the topical focus by defining questions, and then sign a petition committing to active usership once created. These steps are carried out in a forum powered by Stack Overflow's characteristic question-and-answer software.
Outside of Math Overflow, which was the brainchild of a group of Berkeley grad students and post-docs, most of the highly focused results-oriented forums were conceived in this manner. Cartaino told me that the team has created nearly 30 topical sites in the Stack Exchange Network, most of them through the Area 51 staging process. "Physics is the most recent iteration in a long chain of sites," said Cartaino. "It was created because the community asked for it."
The software's structure -- particularly the step-by-step "debugging" of intermediary problems that made it particularly appealing to the team at Math Overflow -- often constrains the applicability of some proposals. "We often get people who like the software and want to use it for purposes the software wasn't intended for," explained Cartaino. "There are topics that aren't well suited to a Stack Exchange site, primarily those that lend themselves more to disucssions. These are meant to be expert communities; question and answer topics work best. Some groups want their own iteration of Stack Exchange for tech support or knowledge management and either aren't aware or aren't interested in the community aspect. We're just not in that market anymore. Everything is sort of created by the community."
But while Stack Exchange software lends itself to more to step-by-step than long, winding discussions, this doesn't confine its usage to more logical disciplines like mathematics or computer programming. Cartaino's team is currently working on a writing-oriented forum. "There's already an Engliush site focusing on the mechanics of writing," said Cartaino. "We're trying to figure out what good questions would look like. 'What is the advantage of writing in the first person?' 'What steps should I take in character development?' 'Is there a good site for start-publishing or selling short runs of hard copy.'" Writing might have a tendency to become more subjective, so we need to stick to the canon: if this is a profession and you guys are experts, what are the canonical questions that need to be asked?"
No matter the focus of each iteration of Stack Overflow's software, each site tends to be successful thanks to Area 51's petition model. "What's really interesting about the Area 51 process is this: there are a lot of ideas for great sites, but Area 51 doesn't create 'em until they have a critical mass of an audience so there's no 'empty restaurant' syndrome," said Cartaino. "It's interesting to see what sites actually get created, and most of the ones that get created are really overnight successes. They don't get launched until there's a community to support it. It takes such a dedicated source of users to get it through the process that they simply have to make it an overnight success."