For the last six weeks or so, Tim Gowers and Terry Tao have been using their blogs (here and here) to carry out a very interesting experiment in large scale collaboration. Gowers chose an open problem and created a system for organizing comments that allowed anyone to contribute to the discussion. Now they seem to have solved the problem and Gowers has posted a blog entry with his thoughts on how the project went and what he learned from the experience. Following the whole project requires a lot of combinatorics background, and would probably require a good amount of time even for an expert (there are over 1000, mostly long, comments in the discussion.) But I think anyone who’s interested in using the web to enhance math research should take a look at his discussion of the experiment.
This collaboration turned out to not be as massive as originally hoped, with only a handful of active participants. I think I’m a little too old fashioned to believe that truly massive collaborations could work (though I’d be happy to be proved wrong) but I think there are other models that could. Gowers suggests the idea of having a number of small groups of collaborators, working on difference but related problems, sharing their discussions and feeding off each other. Or, perhaps it would be useful to have a general discussion in order to allow the participants to find collaborators who are interested in the same techniques or problems. What we need to do is try different experiments until we find models that work.
I don’t know if this blog has a big enough reader base to try such an experiment, but perhaps it’s a good time to start brainstorming. (The “Name the mathematical object contest” was pretty successful, but I don’t think that counts as collaboration…) Do any of you readers in the topology community have thoughts about collaborative projects/experiments that might be worth trying? What are the constraints and conditions that would be necessary to make such a project successful? What sorts of outcomes would we look for?