Thursday, April 08, 2004

The EECS department at Berkeley will hire a few new faculty members next year, and so we've had several interesting job talks. Earlier this week, I heard a candidate talk that touched problems from combinatorics, PDE-constrained optimization, and quantum computation -- and was still comprehensible. At least, I thought he was comprehensible. A job talk is an odd balancing act: there's always too much material to cover, but it must be presented in a way that is still accessible and interesting. There's an edge of tension to a job talk that's missing in an ordinary seminar talk. The matrix computations seminar today interested me just as much as the candidate talk, drew from nearly as many disciplines, and similarly balanced the completeness and accessibility of the presentation; but it was much more relaxed, and I think that contributed to my enjoyment and understanding. Besides, today's speaker spoke with a Dutch accent which I found easy to listen to.

I gave a talk yesterday to a group of MEMS engineers. It turned out to be a very small group; everyone who would normally attend was busy, including the guy who originally invited me to speak. Such is life. I figured out some interesting things in the process of preparing the talk, and I may have convinced someone to use my software, so I can hardly complain.

In a way, classes are a pleasant break from preparing and hearing talks, not to mention from coding and writing. We're studying stability in my fluids course. The material so far is elementary, or at least familiar, but I've still remained entertained. How do you tell if your roommate is stable? asked the professor. Wait until he's at equilibrium -- asleep on the sofa, say -- and then make a loud noise to perturb him. If he quickly goes back to sleep, he's very stable. If he runs to the kitchen, grabs a knife, and chases you around the apartment, he's probably unstable.

I also amused myself for half an hour this morning solving a question asked by a colleague: what does this series sum to?

n = 0
sin((2n+1)y) ⁄ (2n+1)

The answer is π/4 -- independent of y. It's simple if you remember the appropriate calculus tricks. I'm still unsure why Jason cared about this sum, though I assume it was probably for some sort of series solution to a differential equation. Still, I was grateful for the brief distraction. Short, easy problems are a boon. I have spent too much focus on work recently, and it's starting to show; my eating and sleeping patterns are both off.

A fellow student in my graduate algebra class at Maryland once claimed that Gauss went clean-shaven when he worked on algebra, but went scruffy when he did analysis. I expect the story was apocryphal, but if it was not, I seem to be mimicking Gauss in at least one trivial way. Yesterday and today were full of analysis, leading to satisfactory ends in some cases (I have a perturbation-based method for thermoelastic loss calculations which runs way faster than the straightforward approach), and less satisfactory results in other cases (using slightly different versions of the same formula, I get answers which range from 50 to 50000 -- and I trust none of them). Even when I'm not entirely successful, I find this sort of analysis satisfying; unless there's some analysis to aim a computation, the computer will usually give answers which are at best useless and at worst misleading.

However much I admire Gauss and enjoy analysis, I do need to shave.

I've done a poor job of maintaining this blog lately, and have only done an adequate job of responding to friendly e-mails. I have taken some time from work to do other things, though. I walked home a couple times in the last week. The route I take is about five miles, mostly downhill, and it takes me just over an hour and a quarter to walk it. I usually watch the sunset at the end of the walk; I often smell the scents of evening. One day last week, I stopped at a cafe near campus on my walk home. There were several undergrads there, chatting merrily with each other. One pair was discussing fraternity politics; another was talking about which majors on campus were most difficult (with each party in the conversation convinced that he surely had the most difficult major). They spoke loudly enough that I couldn't easily ignore them, and I mentally rolled my eyes a few times as I listened. Was I ever that young?

Of course I was that young. I still am. But in every young man, an old man bides his time. The converse might be true as well, if those who believe in reincarnation have the right of it.

  • Currently drinking: Hot water with lime and honey