I recently (late June, but other things got in the way of this report) attended the Society for Social Choice and Welfare meeting in Boston and the Computational Social Choice meeting in Pittsburgh. Here follows a short report.
SSCW was held at Boston College with around 300 attendees.
Positives: The meeting was overall well organized and the open air conference “clambake” dinner a particular highlight. There were several famous speakers from the Boston area, including Amartya Sen and Daron Acemoglu.
Negatives: Holding parallel sessions in different buildings made switching sessions difficult, which was a problem because there were many sessions that clashed (at least for me). The schedule was very tough, with talks starting at 0800 most days, and the lack of accommodation on campus meant that, with travel time, each day was about 11 hours long. Several of the invited speakers were very hard to track down afterwards, and I suspect they only came in for their talk.
COMSOC 2014 was held at Carnegie Mellon University with around 70 attendees. It was nicely hosted by Ariel Procaccia, and the conference dinner among the dinosaurs at the Carnegie Museum of Natural History was even better than the SSCW one. While less gruelling than the week before, this meeting was certainly a test of stamina for me. Carnegie-Mellon Computer Science was an impressive place, with the architecture of the Gates building and the conference auditorium coming complete with individually sponsored seats including some reserved for Manuel Blum and relatives (how many other university departments worldwide have 3 members of the same family as professors?)
I came away with some new ideas, but not as many as I had hoped, from these two meetings.
Perhaps it is time to rethink conferences of this sort. I would like to see more time for discussion, which probably means reducing the talk length even more. The main purpose of such meetings is to meet people and talk to them, but if you are sitting in talks all the time, this is not easy. Maybe everyone should give a 5-minute (or shorter) talk to advertise, and then have an enormous “poster session” where questions can be asked in detail. Speed dating for researchers has a certain appeal, perhaps.