Xref: utzoo comp.lang.c++:10578 comp.std.c++:451
Path: utzoo!censor!geac!torsqnt!news-server.csri.toronto.edu!cs.utexas.edu!sdd.hp.com!uakari.primate.wisc.edu!uflorida!haven!uvaarpa!murdoch!news
From: randall@Virginia.EDU (Ran Atkinson)
Newsgroups: comp.lang.c++,comp.std.c++
Subject: Re: standard library
Message-ID: <1990Nov26.135542.21825@murdoch.acc.Virginia.EDU>
Date: 26 Nov 90 13:55:42 GMT
References: <1018@zinn.MV.COM>
Sender: news@murdoch.acc.Virginia.EDU
Reply-To: Ran Atkinson <randall@Virginia.EDU>
Followup-To: comp.lang.c++
Organization: University of Virginia
Lines: 24

In article <1018@zinn.MV.COM> mjv@objects.mv.com (Michael J. Vilot) writes:

>Steve Clamage did an independent iostreams implementation for TauMetric.  He is
>helping X3J16 define an I/O library that is implementable.  The standard I/O
>library is based on iostreams (i.e. AT&T 2.0), with suggestions for simplifying
>it from Jerry Schwarz, Doug Lea, and others.  One comment we received last week
>helped us remove the Unix-specific file descriptors (as `int's), so a vendor 
>trying to provide a C++ implementation on an IBM 370 system could do so.

I would hope that the X3J16 committee would be very careful in working
with the standard library to remain 100% POSIX compatible.  The ANSI C
committee worked in close coordination with the POSIX.1 committee to
avoid overlapping work and to ensure that the two standards didn't
conflict unnecessarily.  A rapporteur between X3J16 and POSIX.1 would
be very worthwhile if there isn't one already.

It is worth noting that POSIX is (and has been) fully implementable
on a very wide variety of systems including IBM 370 series and
CRAYs and more conventional ones so being POSIX compatible enhances
portability to a variety of architectures rather than restricting
it.

Ran Atkinson
randall@Virginia.EDU