Drew McCormack wrote:
> This is certainly true of low-level code. What I was talking about was
> higher level abstraction, like representing data structures.
I will not enter this discussion into detail, as it is an old one and a
lost cause, either way. I of course completely agree with Drew here
(seing as how I am one of the J3 folks who worked on OOP stuff, at least
in late stages, and wants interfaces, genericity, etc...). On some
things I agree with J. Giles, inheritance and other "OOP" things are
heavily overused and abused, but this to me says nothing about the
concept itself...C is heavily overused and abused also, and Fortran is
heavily underused and underdeveloped. But C has also produced OSs which
at least work, and C++ has produced things which I use to type this, so
I have no problems with the claim it works for something. And F77 has
produced monster codes which may optimize well but are pure and utter
clutter. If someone writes an e-mail client in Fortran, I will change my
mind...Everything has its purpose, and if only *we* could stick to what
the designers intended and somehow found a way to have the languages
live in harmony...
But the original subject here has a clear answer which was given here
many times, and if a manager cannot see it, quit your job please:
Converting ANY code from Fortran to C is silly, especially if by convert
you don't mean rewrite completely. A code can only gain clarity and
maintainibility if rewritten, otherwise it is merely a syntax change.
And any code that can in fact be converted to C is bound to use such
simple Fortran that it will take 2h to learn for a good programmer. And
if rewriting, choose another language to begin with!
Best,
Aleksandar
P.S. Can we please kick Gerry Tomas off the list? I think such behavior
(even if e-behavior) should not be acceptable on comp.fortran!
|