Path: utzoo!utgpu!news-server.csri.toronto.edu!bonnie.concordia.ca!uunet!cs.utexas.edu!yale!cmcl2!kramden.acf.nyu.edu!brnstnd
From: brnstnd@kramden.acf.nyu.edu (Dan Bernstein)
Newsgroups: comp.lang.misc
Subject: Re: On whether C has first-class composable functions
Message-ID: <1389:Jan905:17:1991@kramden.acf.nyu.edu>
Date: 9 Jan 91 05:17:19 GMT
References: <1990Dec29.110202.3862@mathrt0.math.chalmers.se> <1991Jan5.182428.615@mathrt0.math.chalmers.se> <289@smds.UUCP>
Organization: IR
Lines: 22

In article <289@smds.UUCP> sw@smds.UUCP (Stephen E. Witham) writes:
> To me, the "power" of a language has to do with what concepts you can express
> DIRECTLY.

But what's direct? Why are pointers to functions direct, if pointers to
structures containing blah blah blah are indirect? I think this gives
you the same problem as trying to qualify ``has.''

I prefer to keep ``power'' an objective term, and to qualify it with
further objective terms: memory use, run time, code complexity (by any
sensible measure), etc. Words describing syntax can be objective too,
though I prefer to keep them well separated from the syntactic terms.

  [ function composition routines ]
> The question I'm interested in is, given that you hide the right stuff in
> include files and subroutine libraries, how TWISTED is this method to use
> in the places where you'd want "first class, composable, anonymous functions?"

Looks like it works to me, and it's not sufficiently twisted that I
wouldn't use it.

---Dan