Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!wuarchive!rex!ames!apple!agate!shelby!csli!poser
From: poser@csli.Stanford.EDU (Bill Poser)
Newsgroups: comp.lang.misc
Subject: Re: Fortran vs. C for numerical work - expression notation
Message-ID: <16799@csli.Stanford.EDU>
Date: 9 Dec 90 02:56:19 GMT
References: <16798@csli.Stanford.EDU> <8339@lanl.gov>
Reply-To: poser@csli.stanford.edu (Bill Poser)
Organization: Center for the Study of Language and Information, Stanford U.
Lines: 35

Jim, I hardly came to the "religious defence of C". Rather, I asked that
you give evidence of your claim that Fortran's mathematical
notation is substantially closer to that used in mathematics than C's,
since this had been explicitly discussed. As it turns out, you don't really
examples different from those already mentioned, but apparently you
think that they are more significant.

On this, two comments. First, I don't really accept the claim that
C is further from math than Fortran because it uses, e.g. & for
logical and. If, as you suggest, we were to use /\, we'd still have a
different symbol, one that looks sort of like the mathematical one, but
not quite, and that doesn't type as a single symbol. 

Second, my experience has been that people who study mathematics
at all seriously, both pure and applied, are comfortable with a wide
variety of notations. Look at physics, for example, and the bewildering
variety of subscript and superscript notation you find, and tensor
notation, etc. It is very difficult for me to believe that the difference
between infix and prefix power functions and the like can make much
of a difference to anyone doing serious mathematical computation.
Is there any evidence that it actually matters?

Regarding Piercarlo Grandi's argument that programming notation
should differ as much as possible from mathematical notation,
I am not terribly sympathetic. I suspect that this will just
make people spend their time learning the funny new notation, not
make them think harder about how actual digital computation differs
from symbolic or ideal continuous numerical computation. He could be
right, but I wouldn't bet on it. So, no, I'm not advocating gratuitious
differences, just suggesting: (a) that Fortran and C are not very
different in this respect; (b) that these differences probably don't
matter very much.

							Bill