Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!unmvax!pprg.unm.edu!hc!lanl!jlg
From: jlg@lanl.gov (Jim Giles)
Newsgroups: comp.lang.c
Subject: Re: fortran to C converter
Message-ID: <12661@lanl.gov>
Date: 25 Apr 89 23:29:07 GMT
References: <12787@haddock.ima.isc.com>
Organization: Los Alamos National Laboratory
Lines: 19

From article <12787@haddock.ima.isc.com>, by karl@haddock.ima.isc.com (Karl Heuer):
> [...]               Assuming they have only two floating-point sizes, they
> should have mapped the C type `double' to the smaller one, and used `long
> double' for the less efficient type.

Actually, I think that this is what Cray's C compiler actually does:
it implements both "float" and "double" as the default 64-bit hardware
floats - and there _is_no_other_precision_.  This, of course is almost
as bad as the description I gave in my original posting.  In any case,
with respect to the original subject, whichever way Cray did it: C is
not yet capable of adequately replacing Fortran (I doubt that it ever
will be).

By the way, what _is_ "long double"?  I've never seen a C compiler which
has such a thing.  The proposed ANSI standard mentions "long double",
but doesn't require it to be more precise than double!  In fact, the
only surprising thing about floating point in the standard is that
it requires the FLT_RADIX to be 2 - that is, it requires "float" to
be binary!  No radix is defined for "double" or "long double".