Monday, 8 January 2007

texhacking and a little haskell

Spent this afternoon (still not doing a good job on the whole early-to-bed thing) tweaking BibTeX and LaTeX stuff. It's the kind of job that has the potential to become mindless if you don't pay attention. Useful to pay attentiot to decisions like "well, is this cite more appropriate, or that one? If I cut space consumption this way, will the article stop making sense?" Anyway, glad that first chunk is over with.

Handy tip for finding bibtex errors (kowey-generic users), hunt for "database":
make 2> /dev/null | grep database


Now breaking ground on the derivation output code. Now instead of outputting sentences, I want to output sentences and their derivation. So I have to change (1) the builders (2) the builder interface (3) geni calling the builders (4) the user interface, the whole vertical chain. This kind of work makes me glad for strong/static typing. Sure you try to pay attention, but having the compiler there just telling you "ah no, you missed a spot"... that's very reassuring. It's much better than coding it up, testing it out, and watching it fall on its face (lather, rinse, repeat). The macho programmer might say something like "oh, well if it falls on its face, it's my own fault for not paying attention", but me, I'd rather have the smug annoying friend telling me that I'm headed for disaster, forcing me to get it right before I even think about running that thing.

No comments: