We ended up taping posters to the walls when the stands didn't show, and A ended up sitting with his laptop trying to check people off as they picked up proceedings, but the food was plentiful, and only one talk ran over.
The keynote speaker (David Brooks: Harvard Sensor Networks / HPC Lab) was more than worth the cost of flying him out. He directly talked about the power consumption issues on the motes the sensor net team here is looking at and trying to get into a porcupine-ready state. Also, he made a very good case for the future of processor architecture being centered around the pipeline depth of the chip (based on the current paradigm, chips at higher frequencies will start hitting power consumption points that are on the order of a reactor by 2010, if anyone was fool enough to keep producing in that direction).
Even modern processors do very poorly if you overclock them a bit and remove the heatsink: http://video.google.com/videoplay?docid=5393904704265757054 (<2 minute video, the moral of which is that there are points in life where wooden tables are just not appropriate.)