nawl

My visit to Princeton left me in much the same confused state as before I visited anything.

It was a rather blah day, gray and hazy without any of the niceness of fog, when I arrived on Sunday.

On Monday, I visited the IAS for their morning seminar, which was on game theory on networks, and aimed at an elementary enough level that I actually stayed awake and understood almost everything. The audience asked a lot of questions. I was impressed. The conversation over lunch in the IAS cafeteria centered around a compare and contrast session of admissions and recruitment policies in CS theory before directing itself to removal policies for bad students and the horror stories of the advisor who didn't recognize his student and the student who was disgruntled with his advisor and so killed a random professor with an axe. Awesome recruitment strategy.

So the downside to coming on a random day is apparently that, well, nobody is around. I only got to talk to three professors, because everyone else anybody could think of was out of town or not responding to email. At least I got to have really long conversations, and I learned about some new things like zero knowledge proofs, derandomization techniques, and approximation algorithm techniques. After a certain point it seemed that every time I went in search of direction for a new professor to talk to I was shuttled off to talk to some more grad students.

So while everyone was nice enough and the grad students professed to be happy enough, I was sort of feeling my decision being made for me. Until coffee (rather, chai) the next day with a Slovakian grad student who had randomly been in the restaurant the kids from the department who had been assigned to babysit me took me to the night before. (In a small world note, it turns out he is from the same small town as Brano, the one Slovakian we met in Bratislava a few weeks ago.) Later in the evening we met at the d-bar in the basement of the grad college for a rather amusing, smoky, and French-accented conversation.

So, um, compare and contrast:

Research: Both places are fine. I don't know exactly who I'd end up working with, but it seems like there are a few possibilities at each place. The professors who actually talked to me at Princeton are much younger than the average professor at CMU, so it felt like their interests were much more specific. Princeton seems to have a lot of the approximation algorithms thing going on, which I'm not sure is really my thing. If Steven Rudich at CMU is really so deep into the consciousness thing, there's not much choice for complexity. Since I don't yet know exactly what I want to do, it's probably mostly a matter of me being shifted one way or another by the focus of the department.

Graduation requirements: Almost identical. 6 required classes, 2 semesters TA. Princeton has a qualifying thingy which almost everyone passes. CMU's classes are pass/no pass. Average time to graduation is 5.5 or 6 years. CMU has a lot of department structure to keep track of students. Princeton is small enough that they claim everyone knows about your progress. Funding isn't supposed to be an issue in either place.

Grad students: I was more impressed overall with the CMU kids, and the department appeared to have more social cohesion, even correcting for the visit day coming together thing. But Princeton has a lot of awesome departments, and the most interesting Princeton kids took their friends from all over. Thing #1 that impressed me was that this actually happened. Relief. Dreams of baby bab5. If I went to CMU, I'd probably spend all my time with CS people.

Stuff to do: Pittsburgh is a small but livable city. Princeton is, well, kinda boring and expensive, but as they say, New York is only an hour away. Thing #2 that impressed me was that there were people who did actually go there every weekend. I guess it comes out about equal. Both campuses have ballroom clubs but it's not clear exactly how active each is.

Size: There are ~400 students in SCS at CMU, ~200 in CS, and 25 in theory. There are ~2000 grad students at Princeton, ~100 in CS, and ~20 in theory. Last year there was one new theory student at Princeton. There is one girl and one american among the theory students at Princeton. There are more than that at CMU.

Housing: Everyone lives in their own houses or apartments off campus at CMU. At Princeton, nearly everyone lives in on-campus housing. I was advised to live in the grad college to meet people. The old buildings look like a castle. The dining hall looks like Hogwarts. They have some sort of complicated housing lottery system. The grad college is separated from the rest of the Princeton campus, apparently the result of some university official who thought that grad students should be living a monastic scholarly life.

Prestige: Non-CS people say "Are you kidding? How could you turn down Princeton?" CS people say "Are you kidding? How could you turn down CMU?" Theory people say, um, conflicting things.

So at this point, I guess the responsible thing to do is think long and hard about exactly what I want to do and exactly who I would have as an advisor. But that's less fun than making an informal poll.

So... where should I go?