Xref: utzoo comp.lang.fortran:4774 sci.math:15154 comp.edu:4015
Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uunet!brunix!cs.brown.edu!thc
From: thc@cs.brown.edu (Thomas Colthurst)
Newsgroups: comp.lang.fortran,sci.math,comp.edu
Subject: Re: Fortran for Math Majors
Message-ID: <65195@brunix.UUCP>
Date: 15 Feb 91 23:05:24 GMT
References: <1991Feb12.151552.17331@pdn.paradyne.com> <13138@hubcap.clemson.edu>
Sender: news@brunix.UUCP
Reply-To: thc@cs.brown.edu (Thomas Colthurst)
Organization: Brown Computer Science Dept.
Lines: 20

Most mathematicians, in my experience, have no need for Fortran, and get by rather
well with Mathematica, Maple, MATLAB, etc.  Applied mathematicians, on the other hand,
frequently need to know FORTRAN because of the large amount of FORTRAN code that is
already out there, and because of its wide-spread support (As the saying goes, FORTRAN
is still the language of super-computers.)

I would buy a PC version of FORTRAN only if I know a large deal of my assignments would
be in FORTRAN (which seems unlikely), and even then a modem and good terminal emulator
would seem more useful.

The above statements in no way imply that FORTRAN is the language which applied 
mathematicians should be using.  This honor probably goes to C++, in which it is easy to
implement and use mathematical objects using classes (most come with complex numbers,
vectors, matrices, etc.), is reasonably fast (as good as C, if done right), and 
reasonably portable (anywhere you have a C compiler and can port cfront too).

Pretty soon applied mathematicians are going to start tackling large software projects
and discover that the techniques and languages they are using are totally inappropriate.

-Thomas C