Tuesday, January 20, 2004

First day of classes for Spring 2004! The first fluid mechanics lecture is this afternoon; the starting topic for the course seems to be flow at very low Reynolds numbers. Fun stuff.

I bought a bookshelf and an under-bed plastic storage unit at Target on Saturday. I carried it home, assembled it, put books on it, and suddenly found that I could see the spines on most of my books! But there was still some double-stacking, and if I put one more book on my desk I would probably reach critical mass and the whole pile would self-destruct. So after we returned from a dance lesson at Stanford, I asked Winnie if she would drive me to Target for a second set of shelves. She was willing, and getting the shelves home proved much easier with a car. So on Sunday night, I assembled another set of shelves, reshelved the forty books on my desk (I counted) and shelved the remaining double-stacked books. I filled up two of the three shelves in the unit, so I have some room for expansion.

Whee! Books! I found a few that I thought I'd lost or loaned away. Hello, Mythical Man Month! Fancy meeting you here.

Yes, I talk to my books. No, they don't talk back.

I spent the morning of MLK day reading prognostications about a digital Pearl Harbor, stupid patent tricks, and a college kid named Mike Rowe who set up a site called MikeRoweSoft -- and was sued by Microsoft for it. I came away growling. Computer security experts whose sole qualifications are in management of a security-related company, or in dire prognostications, can do an awful lot to improperly alarm a credulous public. Is computer security important? Yes. Do these authors form plausible scenarios, or write about countermeasures that make sense? Well, some of them are probably the same people who think it's vitally important that public restrooms in the BART train stations be unavailable in order to thwart terrorists (who presumably will be so distracted by full bladders that they'll become confused and head home).

I spent the afternoon and evening becoming equally irritated by linker errors on Solaris. I fixed my interface so it should now actually be portable. But I don't think the thirty pages of linker warnings have anything to do with my code being correct or not; I think they have something to do with the linker receiving the wrong options. I could be wrong, though. The error messages are singularly unhelpful, and the code takes long enough to compile that trial-and-error is trying as it is error-prone.