Path: utzoo!attcan!uunet!cs.utexas.edu!usc!ucsd!hub.ucsb.edu!ucsbuxa!3003jalp
From: 3003jalp@ucsbuxa.ucsb.edu (Applied Magnetics)
Newsgroups: comp.lang.fortran
Subject: Re: C vs. Fortran
Message-ID: <7826@hub.ucsb.edu>
Date: 13 Dec 90 17:02:47 GMT
References: <276765ee@ThreeL.co.uk>
Sender: news@hub.ucsb.edu
Lines: 32

In article <276765ee@ThreeL.co.uk> jf@threel.co.uk (John Fisher) writes:

>ttw@lanl.gov (Tony Warnock) remarks, in an aside:

>[...]
>The discussion seems to me to get bogged down in two rather laughable
>attitudes:
>-- The macho-macho Fortran "you guys don't understand the real world"
>   stance exemplified by the 50 trillion element array man.
>-- The "preserve the purity of the truth" C people, like the person who,
>   when asked for an introduction to C for engineers, simply replied that
>   engineering problems wouldn't provide a complete understanding of
>   C's facilities.
>[...]

I (re-) submit the following examples of possible use for C:
  -Finite elements with random-ish triangular or tetrahedral meshes.
  -Sparse matrices.
Regarding the latter, see e.g. "Sparse Matrix Computations", edited by
J.R. Bunch and D.J. Rose, Academic Press 1976.  (Probably dated, but
relevant to this thread).  The book is full of graph-theoretical
techniques for sparse Gaussian elimination.  _That_ would be easier to
do in C than in Fortran.

I vaguely remember some work by Rob Pike (then at the Caltech Physics
Department) about percolation in a 2-D grid, ca. 1978.  It involved
sparse graphs and was done in C because it was easier that way.

IMHO, the decision to switch languages is tied to the decision to
switch algorithms.  More examples, anyone?  hello?

  --P. Asselin, R&D, Applied Magnetics Corp.  I speak for me.