Saturday, January 31, 2004

It's a beautiful day. We opened the windows to the apartment earlier, and I can hear the sounds of the wind rustling the branches outside. I plan to go walking with Winnie in about an hour when she arrives in El Cerrito. For the moment, though, I need to stay near the phone -- so it seems like a good time to catch up on some writing.

I already spent some time outside today. My bike is in working order again at last, and I took a ride around the block to try it out. Traffic aside, it was grand to be out riding again. Before the bike ride, I walked to the post office to drop of some mail, and from there to Barnes and Noble for a morning cup of coffee and some browsing. I deliberately left my glasses at home, a tactic which I find usually keeps me from becoming so engrossed in browsing as I otherwise might. Glasses or not, though, I was distracted today. I have two new books in my collection: The Pig Who Sang to the Moon by Jeffrey Masson, and The Value of Science, a collection of three books by the famous French mathematician Henri Poincaré.

Sadly, Poincaré is not as well-known outside mathematics as he deserves to be. He was a giant in mathematics, a physicist of no mean skill, and an author of several popular science articles for the general public. The article cited above (from an excellent site on the history of mathematics) ends with the following quote from Poincaré's funeral, which I think is quite well-put:

[M Poincaré was] a mathematician, geometer, philosopher, and man of letters, who was a kind of poet of the infinite, a kind of bard of science.

We need more such figures today (and not just in Georgia).

My proudest accomplishment of the week is, alas, probably an insignificant one. On Thursday night, I finished writing (rewriting, actually) a small program to print mathematical documents typeset by Kahan for an Epson FX80 printer using a custom font. My program generates Postscript code, which I then usually feed through ps2pdf to obtain a PDF document. Unlike previous versions, this code correctly handles all of the documents which I've received from Kahan over the past couple years, including switching back and forth between fonts. And if the code ever does encounter a special character that I forgot about, it will fail gracefully -- an error message describing the unknown character will be sent to the console, and a placeholder will go into the document.

It may not be the most marvelous thing I've ever done, but it was a lot of fun to write.

If I had any doubts that the spring semester is well and truly underway, they would be quashed after a glance at the seminar schedule this week. The matrix computations seminar began this Thursday with a special seminar -- Gene Golub spoke about the solution of real positive nonsymmetric linear systems -- and there was a guest from Japan who spoke on Friday about some of the work done at Chuo University on combining measured data with finite element models in order to more accurately predict microsystem behavior. I attended both those meetings, but I missed both the presentation by MEMSCAP representatives (MEMSCAP runs several standard MEMS fabrication processes) and the presentations for Sensor Nets Day. The seminar calendar is full of other presentations, too -- they might be less relevant to me, but some sound interesting nonetheless.

One of those interesting presentations is coming soon, but remains to be scheduled. My fluids professor is hosting a visitor this semester who is, among other things, an expert in creeping flows and lubrication theory. We've been talking for the last two weeks about creeping flows. Creeping flows are flows in which viscous effects dominate inertial effects. Think about trying to move a marble through cold honey, by way of example -- unless the marble is fired into the honey at high speeds or something similarly implausible, you can neglect the inertial forces (the ma term in F = ma). Since inertial forces scale with volume, it's usually possible to neglect inertia when dealing with very small objects like single-celled animals or microsystems. Since I spend a lot of time thinking about micron-scale machines, creeping flows are particularly interesting to me. Lubrication theory deals with creeping flows in narrowly confined channels -- like the layer of air between a transparency and a projector, the layer of air that supports an air hockey puck, or even a layer of drying paint. A lot of the flows in MEMS involve creeping flows in a thin film of fluid, so lubrication theory is pretty relevant to me, too. So I'm looking forward to this presentation, even though I don't know when it is.

Of course, the fact that it is relevant to my research is an excuse, if a true excuse. The real reason that I'm going is because I think the topic is intrinsically interesting, and because it sounds as though the presenter is one of the grand old statesmen of fluid mechanics -- and such gentlemen often have an entertaining presentation style.

Of course, going to lectures costs time and attention. After the presentation on Friday by our guest from Japan, I and some of my colleagues presented on MEMS simulation, design, measurement, and model verification in which we've been involved. I'd heard most of the presentations before, but felt obliged to stay for the sake of politeness. Polite or not, though, my concentration was spent halfway through the second mini-presentation. I'd have probably felt miserable and started pining for a coffee break if I hadn't had my pencil and pad with me. As it is, I doodled my way through some calculations which I think may prove useful. I need to sanity-check them tonight or tomorrow in order to be sure.