o The AIRWEB ’05 workshop, with some nice search-engine spam papers. As I said to Brian Davison, the organizer, Lehigh University has frankly never really been on my mental map before, but all of a sudden there’s a lot of activity coming out of Lehigh.
o Wandering around the Ginza area of Tokyo with Jan Pedersen, in search of a particular kind of cracker.
o Rohit’s “Web Services Considered Harmful” panel (with Tim Bray, Adam Bosworth, Mark Baker, and a guy representing Jeff Barr from Amazon). The funny thing about this was that there was no controversy, or even substantial disagreement (all participants and most questioners being RESTafarians), yet it was still good. Rohit moderates extremely well. The prediction of the panelists about what the panel would be like if it reconvened three years from now was: there will be no need for this panel three years from now.
o Seeing a lot of people that I’d been aware of for years, but had never met (like Andrei Broder).
o A lowlight (feeling like a bad host): insistently dragging Tantek, Rohit, and Eric Meyer along with us to our favorite top-of-the-hotel bar very late in the evening, finding it full, and ending up in a dank dark basement bar of another hotel, which offered a few bottles of liquor on the bar and an “all-you-can-drink” special (always a bad sign). (The party died at that point.)
o The Yahoo! party, complete with Yahoo!tinis, good food, and great conversations with more people I’d never met in person (Tim Bray, Mark Manasse).
o Eric Brewer‘s keynote on infotech for developing regions. If you ever have a chance to see this talk, do it — it’s very inspiring. The argument is that you can do Ph.D.-level computer science research _and_ significantly help millions of people at the same time. An example: sensor networks along 50,000 kilometers of river sending data back to software that decides on the optimal strategies for spraying to kill blackfly larva. Result: 30M people protected from disease, 100K square miles of farmland made usable. Brewer argued that research like this is slightly marginalized in peer-reviewed computer science, and that just reversing that marginalization could have a huge impact.