Sunday, February 22, 2004
tokens, snow, and ink
ran out of tokens for utordial on thursday evening. still haven't arranged to get more, which means for the moment i'm disconnected. and then of course, i've been in pittsburg township for two days. talked a lot with paul about scheme, lambda calculus in general, and macros.
anyhow, the cs labs are a crappy (but surprisingly packed) place to be on a sunny afternoon, so this'll be short. (1) syntactic closures are fun, but when trying to work with i realized it's really hard to decided what bits of an expression need to be closed without looking at the semantics of the expression, and even then parameters can still be incorrectly captured. (2) renaming (explicitly: in such a way that the user could never do, or implicitly: like the serial code notation) seems like the most promising way to go about. well, i tried it out on several examples using odd hacks to introduce the new lexical scoping rules. works on paper. started implementing it last night (i'm not trying to do the whole define-syntax thing, just to handle a single pattern and an template).
0 Comments:
