Monday, January 23, 2006

SODA Day II...

Today's invited talk was by Princeton's Larry Peterson. He spoke about the aging of the Internet, and how it can't quite cater to the demands of today's world. One of the challenges of adapting the network is the legacy problem: as he put it, it's hard to convince Cisco to change the interpretation of one bit in their routers.

One of the issues that seems critically important in the internet of today is the issue of control. The original internet grew "under the radar". Developed primarily by researchers and technologists, its governing principles were constrained by technical, rather than political, considerations. The next generation of the internet will not have this luxury, and the brouhaha over ICANN is a tiny indicator of the kinds of battles likely looming. Interesting times...

Jeff Erickson did an admirably thorough stream-of-consciousness rendering of the business meeting, which I won't try to duplicate. There was no beer; absolutely none; nada; zilch; not a drop. Everything seemed pointless and arid after that....

I will say this: I continue to be amazed at the capacity of the theory community to repeat the same arguments on the same topics year after year, while presenting said arguments with the kind of gravitas that suggests deep thought and contemplation. Maybe that's how we write so many papers. This year's topic du jour was the always-delightful "submission size formatting" discussion (one of these days I should make separate pages for each of these annoying arguments, debunking all the annoying comments that everyone makes).

Am I irritated ? of course not; my mouth is strangely dry though. Oh yeah, there was no beer.

It was interesting to see Cliff Stein's list of the most active topics at this conference; there were the usual suspects like approximation algorithms, graph algorithms, and computational geometry, but game theory made a (somewhat) surprising appearance in the top five. In retrospect, this is not surprising, especially when we consider things like the minimax theorem, and the fact that most of theoretical computer science consists of games played against an (often unbounded) adversary. The last few years have seen a steady increase in the number of game theory papers in STOC/FOCS/SODA, and this trend will likely continue.


Disqus for The Geomblog