Path: utzoo!attcan!uunet!lll-winken!arisia!quintus!ok
From: ok@quintus.uucp (Richard A. O'Keefe)
Newsgroups: comp.lang.fortran
Subject: Re: Fortran vs C for computations
Message-ID: <373@quintus.UUCP>
Date: 9 Sep 88 09:52:02 GMT
References: <368@quintus.UUCP> <3341@lanl.gov>
Sender: news@quintus.UUCP
Reply-To: ok@quintus.UUCP (Richard A. O'Keefe)
Organization: Quintus Computer Systems, Inc.
Lines: 33

In article <3341@lanl.gov> jlg@lanl.gov (Jim Giles) writes:
>From article <368@quintus.UUCP>, by ok@quintus.uucp (Richard A. O'Keefe):
>> Sorry, this proof won't go through.  [counterexamples]
>
>You are mixing the hypothetical with specific cases.

But it's precisely the specific EXISTING cases which determine
what is CURRENTLY portable or not.  If it is fair enough to cite
nasty machines to support a claim that languages should not have
floating-point exception handling as a standard part, then it is
fair enough for me to cite existing Fortran and C libraries to
support a claim that what C and Fortran can (*now*) do "portably"
"with the aid libraries" is not the same.

I think that ADA shows the way: the language has a standard means
of catching exceptions *IF* they are signalled and a standard means
of suppressing them *IF* they can be suppressed.

The present situation, where I have access to about a dozen different
kinds of machines (all of which support both Fortran and C), and all
of which are capable of detecting and reporting floating-point exceptions,
but there is no single thing I can put in a Fortran program to exploit
this common capability, is *not* acceptable.

We could have something like
	CALL HANDLE(exception, handler, oldhandler, ierror)
where ierror would be set to a suitable value if the implementation was
not able to catch that kind of error.  For example,
	CALL HANDLE('OVERFLOW', MYOVHD, OLDHD, IERROR)
	IF (IERROR .NE. 0) STOP 'CAN''T CATCH OVERFLOWS'

Good heavens, if we were to accept the "some machines don't want to do it"
argument, we wouldn't have floating-point arithmetic, and then where would
Fortran be?