att & osf

[Richard A. O'Keefe]

In article <644 at picuxa.UUCP> tgr at picuxa.UUCP (Dr. Emilio Lizardo) writes:
>In article <268 at quintus.UUCP> ok at quintus.UUCP (Richard A. O'Keefe) writes:
>:POSIX will be better for _everybody_
>:(except AT&T) than the SVID, and a document which is totally controlled
>:by one company can only be a stopgap as a standard.
>
>Correct me if I'm wrong (as if I need to say that in this group :-) but
>hasn't AT&T publicly committed its support to the POSIX effort?

The SVID itself says "The SVID is consistent with the trial-use standard
(Novemeber 1985), with several minor exceptions.  Full conformance to
the IEEE standard will be strongly considered after its formal approval."
-- The minor exceptions, by the way, are things like 0 -vs- -1 for
non-blocking reads, whether group is inherited from directory or process,
all those little things that always made BSD<->ATT porting such fun.
When the SVID becomes an extension of POSIX, it will then be seen to have
been a stopgap >as a standard<.  It will continue to be extremely useful
as a guide to System V features which go beyond POSIX.

>Why are SVID and POSIX seen to be mutually exclusive?

Who said they were?  I didn't!  Currently, SVID is more "real" than POSIX,
in the not-too-distant future POSIX will be more "standard" than SVID.
The point is that P1003 is going to supplant SVID _as a standard_.

>If you assume that,
>because POSIX will be under the "control" of a group other than AT&T,
>AT&T will suffer, I think you're mistaken.

I didn't say that either.  What I said was that POSIX will be _better_
for everybody (except AT&T).  It will be good for AT&T, but it won't
be better for them than the SVID has been.

>As to the last phrase in the cited text above -- I would hardly call SNA
>"a stopgap as a standard" (although it's not a document per se).

I would hardly call MVS a stopgap as a standard, although it is not a
document per se.  I would hardly call the Statue of Liberty a stopgap
as a standard, although it is not a document per se.  It's a funny thing,
but I'm using Ethernet, not SNA, and you can get Ethernet cards and
drivers for V.3.  And the V.3 "TLI" library (which I cannot claim to
understand) is described in terms of the ISORM, not in SNA terms.

A parallel from the IBM world which is more appropriate:  for quite a
while the IBM "F-level" PL/I compiler was the de facto standard for PL/I.
It hasn't been for some years, and PL/I compiler writers can appeal to
a standard (albeit not the world's most comprehensible document...)
instead of having to imitate whatever one manufacturer chooses to do.

It will be a considerable surprise if the second version of the POSIX
standard won't borrow a lot from AT&T's then current version of the SVID.
The great thing is that standards don't change fast, so we'll have had a
chance to see how AT&T's ideas work out in practice before requiring that
everyone support them.  To give you an example of why this is good:
termcap became a de facto standard.  But AT&T decided to produce terminfo
instead.  For some years I thought of this as NIH syndrome, but it finally
penetrated my skull that terminfo is technically better.