(SC22WG14.323) Re: reaction to Variable Length Array Proposal <9302112209.AA18316acurley.osf.org>

[uunet!netcom.com!segfault!rfg]

  |   
  | James A. Crotinger writes:
  | 
  |     I think you're missing the point. People want to convert their
  |   libraries to C while retaining the same API so as not to break
  |   { existing } customer applications { written in FORTRAN }.
  | 
  | Who wants to do that???
  | 
  | What possible purpose would it serve to take a perfectly good, fully
  | debugged, time-tested set of FORTRAN library routines and simply rewrite
  | them in C?
  
  Obviously Ron, you don't bother by using libc each time you program,
  you instead cons every thing up from raw machine instructions :-)

Thanks Mike for underscoring my point for me.

You are correct that I don't want to rewrite libc.  I shouldn't have to.
It's there.  It works.  I call it.

Likewise for existing FORTRAN libraries.  They are there (today).  They
work.  Why not just call them?  Why do people seem to feel this strange
urge to rewrite those libraries from scratch in some dialect of C when
it is perfectly possible to just call them from C as C is now defined?

I haven't heard anybody complaining that there is any particular problem
(on the C side) when calling FORTRAN routines from C.  (And indeed, I
don't believe that there is any problem there.)  You can call your
FORTRAN library routines from C today!

But that's irrelevant to the suggestions that have been made that we need
to mangle C in order to make it just slightly more user-friendly to call
*new* C library routines *from* dusty deck FORTRAN applications without
having to modify the FORTRAN code by one iota.

Can we at least agree that there are two separate parts to this "inter-
operability" issue, i.e. calling FORTRAN from C and calling C from FORTRAN?

Can we at least agree that calling FORTRAN libraries from C code is not
a problem, and that it never has been?

  This whole debate seems to be mired in an Ivory tower somewhere.

Once again, my friend Mike and I are in agreement.  Some folks seem to want
to argue that it will be esthetically "better" somehow if we provide some
way in which new C code can be called in exactly the same way as some old
dusty deck FORTRAN routine.  I however have tried to look at the issue
from a more pragmatic standpoint, and I've noted that such a capability,
if provided, would have little if any practical benefits in practice.


// Ronald F. Guilmette
//    domain address:	rfgasegfault.uucp
//    uucp address:	...!uunet!netcom.com!segfault!rfg