Standards Update, IEEE 1003.0: POSIX Guide

[Jeffrey S. Haemer]

From:  Jeffrey S. Haemer <jsh at usenix.org>


           An Update on UNIX*-Related Standards Activities

                             August, 1990

                 USENIX Standards Watchdog Committee

          Jeffrey S. Haemer <jsh at usenix.org>, Report Editor

IEEE 1003.0: POSIX Guide

Kevin Lewis <klewis at gucci.dco.dec.com> reports on the July 16-20
meeting in Danvers, Massachusetts:

Dot Zero's rite of passage

For the first time in plenary, the group walked through the entire
guide (221 pages), fine-tuning verbiage.  This walk-through takes Dot
Zero across a threshold: instead of soliciting content to fill up
empty sections, we are now filtering the text we have.  I'm proud
we've gotten this far.  I remember when we started this journey,
virtually from scratch.

By the time we finished the walk-through, we had found that we needed
more structure and parameters: rules to make our walk-throughs more
productive.  I ended my last report with the statement, ``let's see if
we have the self-discipline to get there.'' Here is where some of that
self-discipline comes in...  we'll see at the next meeting who abides
by the rules we have agreed upon.

New Volunteers for OS and UI Sections

Two other good things happened in Danvers.  Tricia Oberndorf is now in
charge of the operating system section of the guide.  Tricia is
project leader for the Navy's Next Generation Computer Resources
Operating System Software Working Group (whew!), which has chosen
POSIX as its base standard.  Heretofore, Jim Isaak had been the
section leader.  Now that he has greater duties to fulfill, as part of
the TCOS ruling class.  Tricia has graciously stepped forward to fill
his shoes.

In addition to this noble deed, Martha (``Marti'') Sczcur (pronounced
``seezur''), from NASA, and Ruth Klein, from AT&T, have picked up the
user interface section, which, up until the April, Utah meeting, had
lain untouched for almost two years.  These are welcome resources.
Both of these welcome volunteers made significant contributions, to
the user-interface section of the recently published draft 8 --

__________

  * UNIXTM is a Registered Trademark of UNIX System Laboratories in
    the United States and other countries.

August, 1990 Standards Update                 IEEE 1003.0: POSIX Guide


				- 2 -

 contributions woefully lacking in draft 7,

What Will We Cut and What's a Profile?

Toward the end of my last report, I stated that Dot Zero still faced
hard decisions in two areas: guide content and profiles.  I think
guide content questions will resolve themselves as we move toward the
mock ballot.  Deadlines, like moving your household, have a tendency
to make you throw away stuff that you otherwise might have kept.
Given our goal of an early 1991 mock ballot, I think we will see a
change in our ``pack rat'' mentality.

You might be wondering what might find itself on the editing-room
floor.  I can offer two sections: Data Interchange and Graphics, whose
demise might come about due to a lack of interest by anyone in the
committee to contribute to them and move them along.  There also seems
to be a lot of redundancy.  Good examples of this are the sections I
am responsible for: Introduction and Scope.  The guide seems to say
the same thing in each of these sections but struggles to make it
sound different.  The fine tuning efforts will root this out.

We're still debating profiles, but a consensus is forming around the
term POSIX profile.  Dot Zero agrees we must define such a profile,
but its elements still elude us.  (This gets into the debate about
whether a ``true'' POSIX profile needs to include 1003.1.  Right now,
there is only one POSIX standard, and it would seem to make sense that
a POSIX profile should include it.  However, there are convincing
arguments to the contrary, such as a profile that specifies 1003.2
(shell and tools) compliance on DOS machines, which cannot support
1003.1.  I think POSIX profiles should include some POSIX standard,
but not any specific one.)_ Also, should Dot Zero make mandatory rules
for profile writers, or just offer basic guidelines?  These two topics
will serve as the focus for much of our discussion in the October,
Seattle meeting.

For uniform resolution of our debates about profiles, we will meet and
coordinate with representatives of the other working groups,
particularly the profile groups.  (Right now, that's real-time,
supercomputing, multiprocessing, and transaction processing.) This
will also help ensure that we hear all issues and key points of view.
The primary debate here focuses on whether Dot Zero should attempt to
put ``teeth'' into the guide.  Does Dot Zero, because of its goal in
providing guidance to profile writers, have any say about the
legitimacy of current or future profiling efforts?  How extensive
should this guidance be?  How does Dot Zero provide guidance in an
area where it lacks technical expertise?  These kinds of questions
frame the debate.  [Editor: What do you think the answers are to these
questions?  Speak up.  If you don't go, and don't have anyone else to
tell, at least tell Kevin.]

August, 1990 Standards Update                 IEEE 1003.0: POSIX Guide


Volume-Number: Volume 21, Number 49