Small Languages
2007-01-27 01:36:03.128778+00 by Dan Lyke 1 comments
Over at the oddly named Boston Diaries, Sean has some interesting musings on small languages. I've been working in C++ a lot recently, and immersing myself in the language and templates and some of the "generic programming" notions that people, especially the "software patterns" folks, are pushing.
I'm a fan of using the language which best lets me express the idioms that I'm trying to express, and I'm not averse to writing the occasional preprocessor that takes some higher level construct and turns it into code, and however its perhaps for just this reason that I'm not a fan of language toolkits.
The problem is that a toolkit isn't a domain specific language, it's an attempt to be yet another generic language. It tends to derive not from the need to solve a problem, but the desire for some sort of mental masturbation, frobbing some neurons which need to be frobbed.
Does replacing a for(...)
loop with some sort of genericized object and an instantiation of the transform(...)
template make anything more readonable? Not in any application I've seen yet, and I have been known to use that latter form.
So new languages are fine, but let them, like Perl, evolve to the task, rather than writing them with the hopes that a task will find them. Because that way lies grossness, and I'm inundated in way too much gross code as it is.