Usenet 1994 October/usenetsourcesnewsgroupsinfomagicoctober1994disk2.iso/std_unix/archive/text0008.txt

Text File | 1994-03-07 | 5.9 KB | 201 lines

Submitted-by: [e-mail] (Nicholas M. Stoughton)

               USENIX Standards Report Editor

   Nicholas M. Stoughton <[e-mail]>, Report Editor


What Next?


So What Next?

Steady-State Behavior

Three years ago, attendance at POSIX meetings was around 340
people.  At Irvine this January, the number had fallen to
165, and it is expected (hoped?) that it will bottom out at
around 150.

So what's happened?  Where have all the contributors gone?
There's never a simple answer to questions like these, and
there are at least two major influencers at work in the
POSIX attendances.  The first is straight economics.  The
world is in a recession.  If your company is losing
millions, or even billions, of dollars each year, is sending
people to POSIX meetings the best way of spending what money
you do have?  Why not work at a distance, through the ballot
process?

Another key factor in the attendance figures is the type of
work that is happening now.  Three years ago, new standards
development was in full swing.  Now we are settling down to
a steady state.  Approved standards for POSIX.1, POSIX.2,
POSIX.3, POSIX.4 (aka POSIX.1b), POSIX.5 and POSIX.9 are all
there.  Maintenance work on these is now a major focus, and
that takes a different type of person on the working group.
At Irvine, a considerable amount of time was spent dealing
with interpretation requests, chiefly for the most blatantly
UNIX standards, POSIX.1 and POSIX.2.

Do we, the UNIX user community, care?  Let's give up going
to these meetings ... after all, someone else is sure to do
the right thing, aren't they?  It can't be all that hard to
maintain a standard!  Anyway, we don't want standards rammed
down our throats at every opportunity.

Unfortunately, the serpent whose name is invention is lying
there, coiled and ready to strike the moment that someone
stops saying ``But where's the existing practice''!  I've
lost track of how many times Jeff Haemer and I have trotted
that phrase out recently.  In the standards world, practice
really does make perfect.  If standards are to be rammed
down our throats, and that's a whole different discussion,
let them at least be palatable ones.













- 2 -



The rules of interpretation say that the approved standard
should be interpreted as loosely as possible.  If it was
actually wrong, it is up to a later revision to fix the
wording, but no one can complain in the interim.  This will
of course lead to lots of ``Weirdnix'' implementations:
systems that claim POSIX conformance, but are about as far
removed from UNIX as you can imagine!

So it is necessary to keep a a significant presence within
POSIX, attempting to restrain and guide the work occurring.
Existing documents need revision to clarify the wording to
help prevent the worst excesses of Weirdnix.  New POSIX
projects are still being introduced, but at a much reduced
level.  Many of these are not necessarily ``mainstream
UNIX'' things either - an ADA interface to time
synchronisation, Test Methods for POSIX.1b, the Realtime
standard (yes, the number _h_a_s changed, it used to be known
as POSIX.4), and so on.  Nevertheless, new direct UNIX
projects are not unthinkable, and we must be ready to meet
these challenges.

To give you a flavour of what is happening in the
interpretations process, Andrew Josey, the Vice Chair of
Interpretations has put together the following information:

Working   Mailing List                   Current Position            Outstanding
Group                                                                Requests
________________________________________________________________________________
P1003.1   [e-mail]   61 request pre-Irvine, 30      31
                                         addressed & in progress
P2003.1   [e-mail]   17 request, 5 complete         12
P1003.2   [e-mail]   30 requests pre-Irvine          0
P1003.5   [e-mail] 9 requests being processed 1
P1224     [e-mail]      6 requests have been answered  0
P1224.1   [e-mail]    2 requests have been completed 0
P1224.2   [e-mail]    3 requests being processed     0
P1327     [e-mail]      3 requests have been answered  0

In the past couple of months new draft guidelines for PASC
interpretations have been circulated and a BOF session met
at Irvine to discuss them further.

The guidelines attempt to address the issues of timely
response, and also of the scope of the interpretations.
They are now being followed and we hope to see some
improvement in the process over the next six months.

What Do You Care About Standards?

I would like to take this opportunity to solicit your
opinions as to what should appear in this column.  I was
recently invited to submit a series of articles to a










- 3 -



prominent Open System Magazine, but after I had sent in a
couple, I was told ``These are far too POSIX-centric.  What
about some other standards''?

There are certainly several other areas that might be useful
to report on, both in the _d_e _f_a_c_t_o and _d_e _j_u_r_e worlds.  But
rather than trying to read your minds, I'll solicit your
suggestions.  What else would you like to hear about?

Thank You

I would like to take this opportunity to publicly thank
Michael Hannah, a regular contributor to this column over
the years, for all his work with POSIX.9, his wit, and his
enthusiasm.  Michael has been promoted to a new position
within Sandia National Labs, running a 2000-node Intel
Paragon system - I know he has some stories that would
interest SAGE members - and will no longer be in a position
to continue his work with POSIX.  The first article I ever
edited for this column was from Michael, and the ease with
which I was able to work with him persuaded me to take on
the job permanently.  I am sure you will all join me in
wishing him every success in the future.






































Volume-Number: Volume 34, Number 5