jason@hpcnoe.UUCP (Jason Zions) (09/16/86)
In the past two weeks or so, I've seen more than a few people post articles
dumping on the Computer Science subfield of Numerical Methods. This article
is aimed at those people and comments.
Yes, I suppose that if you're willing to postulate that programmers NEVER
use floating point numbers, I'm willing to say that a course in Numerical
methods is unnecessary.
Not every computer science major ends up kernel hacking, writing device
drivers, etc. Some are involved in data acquisition and reduction, some
write payroll software (I'm not talking COBOL here, kiddies; you try
writing one payroll package to handle distributed databases, etc. and so
forth...), some do programs that report statistics.
I write and test data communications software. "No floating point there,"
I here you cry. Well, some of our diagnostics report statistics and per-
centages for errors, accumulated counters, etc. Floating point rears its
ugly head.
"But ANY IDIOT can calculate a percentage!" I'm glad you said that. By the
same token, you might say that any idiot knows how to sum a column of numbers
that may of different orders of magnitude. Unfortunately, this is not so;
I've seen third year CS students code this sort of thing up so they add them
in any order, biggest to smallest, etc.
My conclusion (note that this is an OPINION): if you're ever going to have
anything to do with a floating point number, you NEED to have a good under-
standing of numerical methods. You can get it from a book; you can get it
from a class; or you can get it from chasing down an incredibly obscure bug
in your code that "almost always works..."
For those of you that care about grammar, spelling, and forms of argument,
I admit that perhaps some of my argumentative technique in the above was
inappropriate or just plain brow-beating. Perhaps some of you might be willing
to send me e-mail showing how I could improve it? You see, I am one of those
people who care about grammar, spelling, and argumentative technique.
--
This is not an official statement of Hewlett-Packard Corp., and does not
necessarily reflect the views of HP. It is provided completely without warranty
of any kind. Lawyers take 3d10 damage and roll a saving throw vs. ego attack.
Jason Zions Hewlett-Packard
Colorado Networks Division 3404 E. Harmony Road
Mail Stop 102 Ft. Collins, CO 80525
{ihnp4,seismo,hplabs,gatech}!hpfcdc!hpcnoe!jason or jason@hpcnoe.fc.hp.com