Re: Programming Languages, "to C or not to C, that is the Q."

[Michael Marsh <michael.a.marsh@gmail.com>]

On Tue, 25 Jan 2005 15:18:47 -0600, Steve Block <scblock@ev-15.com> wrote:
> I hope to god people aren't still writing F77 code on a daily basis. I'm
> a fairly experience F90 and F95 coder, but I can barely decipher what is going
> on in most F77 programs. Fixed format F77 code has to be one of the strangest
> things anyone ever forced on programmers (I recognize that someone using
> it for years might think it is easy, but I don't).

When I left grad school, my old experiment was still using F77 for
some parts of the data analysis system.  There were a number of people
unwilling to let go of Fortran, and F77 was what the experiment
standardized on (to the extent that anything was "standard").  It
really isn't *that* bad, though it's a very limiting language,
especially if you officially "ban" certain constructs such as computed
GOTO.  We did allow variable and function names greater than six
characters, though.

I used to be responsible for the older codebase, which was almost
entirely in F77 (about half a million lines written by Physicists,
much of it spaghetti).  When I started, I knew absolutely no Fortran. 
A little over a year later, I could do just about anything I needed to
in it.  I even did a bunch of F77-to-C++ interfacing.  If you think
that's sick, I have SWIG wrappers that allow you to call CERNLIB
routines from perl and python.

-- 
Michael A. Marsh
http://www.umiacs.umd.edu/~mmarsh