August 17th, 2009

beartato phd

(no subject)

Worked on thesis revisions a bit. I didn't know Karl actually wrote up his technique for doing linear and modal logics as a TR, it looks pretty nice. I wonder what's really going on between his approach and mine --- I always hit a snag when I try to do modal logics not simpson style but "tethered", and filtering out individual variable uses doesn't really work for ordered logic, and presumably other logics that have more interfering conditions on variable use --- but certainly the question is up in the air as to precisely which context structures count as "interfering".

Talked to Dinitz when I ran into him at weekly Tea in GHC about his summer research on --- convenient coincidence --- network routing algorithms. He had some neat complexity results about an idealized version of interior BGP. There's a problem that looks like it should be as bad as Σ2 because of the way it universally quantifies over all possible subsets of your routers that might have the best external route, but actually because routes live in a metric space, for some reason the set of classes of different subsets is only linear in the number of routers not the full powerset, so it's just NP after all.

Put in an application for an apartment in philly.

Went to dinner with tom7.