Path: utzoo!attcan!uunet!mailrus!iuvax!rutgers!mcnc!rti!bcw
From: bcw@rti.UUCP (Bruce Wright)
Newsgroups: comp.sys.ibm.pc
Subject: Re: PC FORTRAN compiler?
Summary: Microsoft Fortran
Message-ID: <3417@rti.UUCP>
Date: 12 Jan 90 16:38:12 GMT
References: <6595@ogicse.ogc.edu> <21117@kuhub.cc.ukans.edu>
Distribution: na
Organization: Research Triangle Institute, RTP, NC
Lines: 38

In article <21117@kuhub.cc.ukans.edu>, ARRITT@kuhub.cc.ukans.edu (What a waste it is to lose one's mind.) writes:
> The Microsoft compiler is only a very limited subset of Fortran-77.  
> I tried it and got hundreds of error messages.  Although I don't pretend 
> to be a great (or even good) programmer, the code has run on 11 different 
> systems made by 6 manufacturers, ranging from low-end Microvaxes to a 
> Cray X/MP-48.  So it's Microsoft's problem and not any quirk of the code.  
> 
> Moral:
> don't expect a quick and easy port using the Microsoft Fortran compiler.
> In fairness our Microsoft compiler is a couple of years old.  Maybe the more
> recent ones are better.

The earlier versions of the Microsoft Fortran compiler were pretty awful.
The version that bills itself as the "Microsoft Fortran Optimizing Compiler"
was the first one to be halfway reasonable (1987 according to the copyright
notice in the documentation).  It is pretty much a strict Fortran-77 
compiler, without extensions, but is still an *enormous* improvement on
the earlier compilers (which says a lot about the earlier compilers ...).
Many people's idea of Fortran-77 is a much larger language than the actual
standard - no doubt because many manufacturers implemented quite a few
extensions that were (nearly) mutually compatible but which were not
contained in the standard.  The actual F77 language is *much* smaller than
what the Fortran compiler on a VAX will accept, for example.

Recently (like within the last few months), Microsoft has come out with a
new version of their Fortran compiler.  I haven't gotten it because I am
not doing much with Fortran these days (fortunately), but it is supposed
to implement many of the features from other Fortran-77 implementations
that were not actually part of the F77 standard.  They claim that it is
essentially compatible with the DEC VAX Fortran-77 compiler except for
some of the hardware-dependent features in the DEC compiler.  If you
have the older Microsoft compiler you should be able to upgrade for a
reasonable fee - probably what you would pay for even a cheap Fortran-77
compiler from somewhere else, and maybe less.  If you don't have the older
Microsoft compiler you may be able to do better elsewhere, I haven't looked
at the Fortran market lately. :-)

						Bruce C. Wright