One would think it would be easier.
Sometimes, it is, like for instance when a very good speaker submits a very good talk proposal which has no conflicts with any other.
Most times, it isn't.
There's a guy who submitted four talks and all of them are good (the problem here is to choose which ones to accept). There's three guys who submitted what looks like the same talk (we'll have to ask for details, in these cases). There's a couple of talks we just can't understand what they're about, regardless of how long we stare at the description... :-)
Plus, it's a whole bunch of talks, and we also have other things to do (jobs, etc), so we can't afford the time to start on one end and proceed to the other one without stopping, which means we still haven't been able to look at all talks.
More details soon.
(The preliminary reading turned out to be only partly useful - a huge number of submissions came in on the last day or two before the deadline, right up to and beyond the deadline in fact. Because of this last minute submission pattern, we had spent the previous week contacting a number of paper proposers to see whether they would be open to changes in length, and panel discussions, and such; but with the last minute rush we switched from needing to pad things out to instead having to cut things that we wanted to keep.)
As it happened, we finished up in one day (but it was close).
We basically used the same daily time schedule as last year's Buffalo YAPC: morning and afternoon each contained 2 90 minute sessions separated by a 15 minute break. A 90-minute session could contain an 90 minute talk, or 2 40-minute talks, or 4 20-minute talks. (The 10 minute difference between 2 40-minute talks and one 90-minute talk is a small period to switch from one room to another when they both had talks you wanted to see.)
We blocked out the special items (keynotes, town hall, lightning talks) first since these provided the times when there was only one track, so nothing else could be scheduled at the same time in another room. Next, we set up the major tracks. (Two of the people had gone through the list and made track assignments for all of the proposals. The resulting track names turned out to be remarkably similar - great minds think alike and all that.) Some of the track categories we had were intro/training, perl6, testing, web stuff (a number of sub-tracks), database, CPAN/modules, general programming, non-perl interaction (both glue and full applications/success stories). We decided which tracks should not overlap and which could. We slotted the most important tracks that had to be non-conflicting, fit in the remaining tracks where most appropriate (or rather least objectionable). Within a track, we'd choose amongst duplicate submissions and toss out stuff that didn't catch the interest of any of us as the first level of weeding. The track assignment process filled in much of the available time, we took the rest of the proposals and spread them out on a table and went through them sorting as must-have / perhaps / forget-it, decided where to insert the must-haves, removed the forget-its, and repeated until we ran out of space.
Then we did a coherency check. (We found that we'd missed one instance of the same speaker being in two rooms at the same time. We did some fine tuning for overlap conflicts, including finding that we had a set of 20-40-20 talks in one room at the same time as a 20-20-40 minute set in another.)
One thing that was especially useful was that our speaker submission form required a phone number - we made a number of calls to speakers during that meeting. (Are you willing to let us accept *all* of your talks? Can you fit into 20 minutes?
