guy (04/30/83)
/usr/group (or UniForum; they're going back for a second time to choose a
new name) has a Standards Committee which was formed "to formulate, adopt,
publish, and provide a formal standard specification, based on the UNIX
operating system developed by Bell Laboratories, for a commercial
operating system." The need for such a standard should be obvious;
internally to AT&T, there have been more versions of UNIX than you can
shake a stick at, and several of them (Research V6 and V7, PWB/UNIX 1.0,
USG UNIX 3.0 (System III), and USG UNIX 5.0 (System V)) have been released
publicly. There are various differences between them; the OEM's of UNIX
have further made other changes. Since one of the main advantages of UNIX
is its portability (the first time I moved a program from PDP-11 UNIX to
another machine, I was surprised and amazed; now, of course, it's old hat),
it would be nice if one could minimize the number of changes one has to make
to move programs between different machines running UNIX.
The Standards Committee has representatives from Bell Labs, from "the Berkeley
community" (after all, Bill Joy isn't working at Berkeley any more, he's
working at SUN), from the major UNIX OEM's, and from some of the companies
offering UNIX-compatible/UNIX-like systems. They have produced a draft
standard which has been sent to the membership of /usr/group for review;
the goal is to have the standard adopted and ready for distribution on
November 7, 1983.
I think that there would be an interest in discussing UNIX standards (possibly
including C standards - there's an ANSI committee working on a C standard),
and possibly computer standards in general, in USENET. I propose a newsgroup
called "net.standards" for this. (I also encourage people to participate
in the Standards Committee's work.)
Guy Harris
RLG Corporation
{seismo,mcnc,we13,brl-bmd}!rlgvax!guy