O'pain Software Foundation: (2) Why is it better than AT&T?

[Jim Perry]

In article <7986 at brl-smoke.ARPA> gwyn at brl.arpa (Doug Gwyn (VLD/VMB) <gwyn>) writes:
>In article <503 at bacchus.DEC.COM> price at decwrl.UUCP (Chuck Price) writes:
>
>>If AT&T is so committed to open systems, why doesn't it join OSF?
>
>Assuming AT&T has made a decision not to, which I don't think we know
>at this time, it could well be that they are reluctant to abandon the
>operating system that they've been gradually improving with certain
>long-range goals in mind just to pick up a version that starts way
>behind where AT&T's system currently is.

Ah, but that's exactly the point.  This is the mindset of the owner of
a proprietary operating system.  I was not around when the issue of 
supporting Unix was debated at Apollo, but I would imagine the anti-Unix
argument could have been pretty well summed up as "reluctance to abandon
the operating system that they've been gradually improving with certain
long-range goals in mind just to pick up an OS that starts way behind
where Apollo's system currently is".  As it happens, people who buy 
computers want them to run Unix because it is a standard.  Whether it's 
state-of-the-art or antiquated is not relevant, it's a standard.  Sort
of like FORTRAN IV in the '70's, people who know better sneer at it but
everybody sells it (and devotes a lot of resources to it).  But I digress.

A standard is successful because it is a common base on which everyone
can build, with each player adding extensions suited to their strengths,
be it parallel processing, real-time, whatever.  Unix is a standard because
Bell Labs was not in the commercial computer business and therefore made
it widely available, and many people implemented it on a variety of hardware.  
Now AT&T (along with Sun) is in the position of developing it in the character
of a product rather than a research item.  The argument goes that if AT&T+Sun
not only get to make their proprietary extensions but also get to fold those
extensions back into the base at intervals, this puts them at a considerable
competitive advantage.  Having a 'standard' defined by one (or two) companies'
proprietary system may not be a good idea.  Does everybody scramble to update
their Fortran compilers because IBM releases a new version?  

Disclaimer: I don't speak for anyone but myself, and I'm not so sure I'm
always in complete agreement with what I say.  As it happens, I'm an
iconoclastic upstart who doesn't much care for premature standardization...
but that's not the way the wind blows these days.

Jim Perry   perry at apollo.UUCP