Path: utzoo!attcan!uunet!cme!libes
From: libes@cme.nbs.gov (Don Libes)
Newsgroups: comp.unix.questions
Subject: Re: Do OS's slow down with age?
Message-ID: <836@muffin.cme.nbs.gov>
Date: 16 Jan 89 23:30:26 GMT
References: <209@imspw6.UUCP> <12872@steinmetz.ge.com> <370@siswat.UUCP> <1472@cps3xx.UUCP> <12938@steinmetz.ge.com> <2862@kuhub.cc.ukans.edu> <9360@smoke.BRL.MIL> <2880@kuhub.cc.ukans.edu> <9373@smoke.BRL.MIL>
Reply-To: libes@cme.nbs.gov (Don Libes)
Organization: National Institute of Standards and Technology
Lines: 23

In article <9373@smoke.BRL.MIL> gwyn@brl.arpa (Doug Gwyn (VLD/VMB) <gwyn>) writes:
>Most of my applications port unchanged among several radically
>different UNIX implementations, including:  [long list deleted here]
>
>In this context, developing VMS-dependent applications would
>be utterly laughable.

Our project (a heterogenous distributed data system) uses the same
code on both VMS and UNIX systems.  Sure, there are #ifdefs, but it's
not necessarily that bad.  Indeed, the VMS C compiler has an
incredible number of hacks built in just to support UNIX sources.  For
example, you can use UNIX-style file names, and the C runtimes will
convert them appropriately for VMS.  (It almost always works :-)
Similar hacks exist for UNIX library and system calls.

The most painful thing is that it is impossible to have portable
Makefiles.  VMS make (MMS) uses a different syntax than UNIX make.  Argh!

Somewhere I read that DEC is going to have a POSIX-compatible version
of VMS shortly.  This will make life easier, although it ain't gonna
speed up VMS any.

Don Libes          libes@cme.nbs.gov      ...!uunet!cme-durer!libes