Path: utzoo!censor!geac!torsqnt!news-server.csri.toronto.edu!cs.utexas.edu!usc!julius.cs.uiuc.edu!psuvax1!rutgers!uwvax!picard.cs.wisc.edu!quale
From: quale@picard.cs.wisc.edu (Douglas E. Quale)
Newsgroups: comp.lang.misc
Subject: Re: On whether C has first-class composable functions
Message-ID: <1991Jan19.210006.8112@spool.cs.wisc.edu>
Date: 19 Jan 91 21:00:06 GMT
References: <4408:Jan421:44:3391@kramden.acf.nyu.edu> <1991Jan13.045906.6210@spool.cs.wisc.edu> <301@smds.UUCP>
Sender: quale@picard.cs.wisc.edu
Organization: U of Wisconsin CS Dept
Lines: 43

In article <301@smds.UUCP> sw@smds.UUCP (Stephen E. Witham) writes:
>In article <1991Jan13.045906.6210@spool.cs.wisc.edu>, 
>quale@picard.cs.wisc.edu (Douglas E. Quale) is ostensibly writing about the
>various simulations of composable functions that have been posted here.
>He talks about:
> 1)  The evils of fixed-size tables, as in Kristoffer Eriksson's posting.
> 2)  The evils of non-portability (Mr. Eriksson mentioned doing something
>     non-portable).
> 3)  The problem that if you allocate something, you have to explicitly
>     free it in C.
> 4)  That the "compose" functions seen so far only work on one type of
>     function.  They are not polymorphic.
>Finally, he complains that...
> 5)  "Some people will use a hammer for anything, even driving screws."
>
>Numbers 1 and 2 have to do with Mr. Ericsson's posting, and not, as Mr.
>Quale seems to think, the C language or composable functions.

If you know a portable method implementation in C that removes the
restriction to fixed size tables, post it.  Mr. Eriksson said he *didn't*
know a portable way to fix this problem in C, so it *is* a C language
problem.

>
>Number 5 is nasty.  First of all, it's not all that clumsy, although 
>it's not Scheme.  But mainly, this is something we were specifically
>challenged to do.  Now you're making fun of us for trying??  Also, given
>one language at a time to work with, if I'm driving more nails than screws,
>I'll use the hammer language instead of the screw language.
>

Not all that clumsy is a value judgement that you are entitled to make,
and I suppose then that I'm entitled to chuckle as well.  (The C code
is somewhat larger than (define (compose f g) (lambda (x) (f (g x)))),
and I'd really like to see you prove that the C code works, and the
C implementation does have 69 different arbitrary limits built in that
don't exist in a lisp solution, but it's your call.)

If you have to drive nails and screws, I'd use a language that provides
a full toolkit instead of just a hammer.

-- Doug Quale
quale@picard.cs.wisc.edu