brain cramp

[tiedyedave]

It doesn't work. I forgot about contravariance.

How the fuck did I forget about contravariance?!

Comments

[sleepsong.livejournal.com]

I forgot about the parentheses in my research... Goes to show that I should never learn LISP.

[tiedyedave]

However, there are plenty of other reasons not to learn LISP.

[gwillen.livejournal.com]

True, but your reasons all suck when compared to the glory that is LISP.

[tiedyedave]

It seems to me that there have been two really clean, consistent models of programming so far: the C model and

Stopped right there.

[rjmccall.livejournal.com]

Hmm. I couldn't find a context for that quote, but I can find a find grains of truth in it. C reflects the this is hardware, so you'd better know what you're doing facet of PL design, which remains important for some core systems/high-performance work. The problem is the number of applications which don't belong anywhere near that niche yet are nonetheless written in C or C++. A propos, there's a new buffer overflow in zlib.

Anyway, the essay I read was devoted to defending LISP against the Java model, which has a lot of value as an argument but which doesn't address the interesting questions about LISP, e.g. whether completely dynamic typing and interpretation are really such great ideas. It'd be interesting to see more discussion on that front.

[rjmccall.livejournal.com]

Oh, he briefly discusses ML (as an example of statically-typed languages) in the FAQ. He's got a point, but I'm not as sure mixed types are great in the longer term. Amusingly, of all languages JavaScript is approaching the interesting ideal of a dynamic, exploratory language which lets you append typing however you want later on.

[rjmccall.livejournal.com]

I'm sorry if I'm not cool enough to talk to you! I'll just go die in a gutter!

[tiedyedave]

You're damn right you're not cool enough. My advisor is leaving town in three days, and wants a draft. Even after I finish this proof (which at a normal pace would take a week), I still have a week's worth of paper writing to do in order to clothe the technical results with some kind of relevance.

Which means I have time for cutesy remarks, but not a real get-our-facts-straight discussion. Besides, such a discussion would only result in some kind of admission that you're right. And why would I want to do that (for the bajillionth time)?

[tiedyedave]

Verdict: Go die in a gutter. I will be joining you shortly.

[sleepsong.livejournal.com]

Not apropos contravariance but apropos annoying statistics and research and stuff in general:

Do you use logistic analysis in any of your research? I'm having issues and nobody around here knows the answers and teh intarweb has failed me.