(no subject)

I'm in that horrible stage in semester where you just get perpetually tired. I'm either in the labs, in class, in the library, at work, or working at home. Oh, and there are brief periods of unconsciousness in there, when I get home and collapse in a little heap.

The last few days have been spent footnoting law review articles in the library and unit testing the deslanter for our 440 project. I got a relatively easy article to footnote this time - the reply to Kirby's "Welcome to Law Reviews" by John Gava. Interestingly, Kirby had actually written his article in response to a commentary by Gava on the same topic, so it's more or less a conversation conducted over a period of a few years in the pages of two law journals. The issue was brought up in 1936, when Fred Rodell famously said "There are two things wrong with almost all legal writing. One is its style, the other is its content." This was wholeheartedly endorsed by Gava a few years ago (1997, I think). Kirby, on the other hand, pointed out that while there are undoubtedly many things wrong with the plethora of law reviews that are currently being produced, high quality law reviews (such as the MULR, of course) are of great assistance to judges who are pressed for time. Gava replies rather snappily that judges should be capable of producing great judgements without them and therefore avoid the risk of becoming "activist judges". He also goes off onto a little rant about the damage being done to law schools by the "publish or perish" phenomenoen.

On the 440 front, we are now well and truly into the clean-up of the first increment. The classes I'd been assigned - Deslanter, SymbolCode, GlyphList - are a combination of ridiculously simple data structure classes (SymbolCode, GlyphList) and the very mathematical Deslanter. Actuallly, SymbolCode doesn't do anything. It takes a string and gives you back the same string. Given that a string is a primitive type, both I and the SADD manager fail to see the point. In fact, it's not even in the SADD! But hey, at least it didn't take long to write a test for it. Deslanter is the nasty class that I've spent the last couple of days working with. The problem is that I was assigned to do white box testing on it. Which I'm not even sure that I've done correctly. We're using definition-use testing, which I think is supposed to exercise every path from each definition to its uses. But I think I just did branch testing, where every branch is gone down once... hmm... maybe I should talk to the test manager. But the test case wasn't all that good (Mel, remember call by reference?) and so spent about as much time debugging that as that actual code. (And the code had LOTS of bugs.) Unfortunately, the evil PM who wrote the code is leaving for Japan TOMORROW!! The odds of this code getting any better are not good. Ah well, it's not a critical part of the system. Our HR manager is now being stuck with deputy PM job - taking over while he's away as well as cracking the whip. All is good, I don't have to do it, which I was worried about after the results of the management review came out.

This weekend, there's escrima sticks at karate, I have to do a code inspection, write the GlyphList unit test, and pack my things to move home. Oh, and it's Will's 21st party on Saturday. I also have to read and summarise four books so I can return them to the library. *sigh*

I just want to sleep for a week.