Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site brl-tgr.ARPA
Path: utzoo!watmath!clyde!burl!ulysses!allegra!mit-eddie!genrad!panda!talcott!harvard!seismo!brl-tgr!gwyn
From: gwyn@brl-tgr.ARPA (Doug Gwyn <gwyn>)
Newsgroups: net.lang.c
Subject: Re: AT&T 7300 C compiler
Message-ID: <10016@brl-tgr.ARPA>
Date: Wed, 17-Apr-85 11:36:24 EST
Article-I.D.: brl-tgr.10016
Posted: Wed Apr 17 11:36:24 1985
Date-Received: Fri, 19-Apr-85 00:13:23 EST
References: <9726@brl-tgr.ARPA> <5444@utzoo.UUCP>, <9881@brl-tgr.ARPA> <5477@utzoo.UUCP>
Organization: Ballistic Research Lab
Lines: 15

> More seriously, my original comment was motivated by amusement (and
> some disgust) at AT&T simultaneously pushing System V as a "standard"
> and continuing to change it in incompatible ways.

I haven't noticed any problem tracking the evolution of the AT&T UNIX
product from UNIX System III through UNIX System V Release 2 Version 2.
They have been rather careful to maintain the previous system interface
with each new release (there have been a couple of minor slip-ups but
nothing like the changes from, say, 4.1BSD to 4.2BSD).  Changes have
come in the form of new additions or extensions to existing facilities,
better specifications, and hidden internal improvements.  The System V
Interface Definition provides a controlled way to phase out obsolete
facilities while maintaining a stable system interface for applications.
This looks like a big win to me, whether implemented by AT&T or by other
vendors.