Splinter Unix?

[Chuck Karish]

In article <7932 at brl-smoke.ARPA> gwyn at brl.arpa (Doug Gwyn (VLD/VMB) <gwyn>) writes:
>In article <21387 at labrea.STANFORD.EDU> karish at denali.stanford.edu (Chuck Karish) writes:
>>Is there any indication that OSF intends to write a complete,
>>incompatible implementation?
>
>Is there any doubt that that is what will happen?  Take STREAMS,
>including RFS, for example.  It is rather hard to implement this
>extremely important post-SVR2 feature simply from the non-proprietary
>specifications (at least from those of which I am aware) without
>introducing SOME degree of incompatibility with AT&T-based
>implementations.

Are we still talking about an open standard?

>>... others say that the way they plan to do it will give Sun an
>>unfair marketing advantage (several months) over their competitors
>>(and uneasy bedfellows).
>
>Seems to me the noisy vendors had plenty of time to work out a
>similar deal with AT&T.  Is it unfair for a company that sees a
>need and works to meet it to gain a competitive advantage thereby?

No.  I'm sure that the Sun/AT&T product will be a good one, and they
deserve to profit from their efforts.  Their competitors, however,
can't afford to base their business strategies on what Sun might or
might not deliver to them.  Perhaps my use of the word `unfair' was
misunderstood; I meant it to apply to the metaphorical game I'd just
described.

I think it's tragic that it's impossible to make real improvements to
Unix without making big hunks of it proprietary, but that's the way our
economic system works.

>I think not.  (By the way, I don't know that they really will.)

If they don't, they won't be doing their jobs, and their stockholders
should hold them to account.

>Or is "fair" supposed to mean that companies who haven't contributed
>to the development of UNIX are supposed to parasitically reap rewards
>from it?  They should count themselves lucky that people even buy
>their systems after they spent years attempting to lock customers
>into their proprietary product lines.

IBM and DEC are probably EACH now putting more resources into Un*x
development than is the Sun/AT&T combine.  AT&T hasn't sold cheap UNIX
licenses for altruistic reasons; they get a lot of free engineering
done in universities, by customers who could never have afforded
full-price software.  The development of UNIX as an open system took
place during a period when AT&T was prohibited by law from promoting
it as a proprietary product.

I appreciate the engineering aesthetics and the
ethics of the people who created UNIX, and I hope to see those
traditions carried on.  Those are the values of the engineers, however,
not of the corporations.

>I agree fully with the fellow
>who sees the AIX ploy as an attempt to destroy UNIX as an open system.
>Let's hope they get what they deserve, which is loss of sales to other
>vendors who offer "common UNIX" with value added.

What the heck is "common UNIX"?  Note that the phrase contains a
trademark owned by AT&T.  Getting back to my earlier point, any element
of an open standard must be reproducible from publicly available,
written specifications.  It's possible to do NFS that way.  It's
possible to do PostScript that way.  If AT&T wants to keep streams as a
proprietary extension, fine.  The customers will decide whether it's a
critical selling point.

You may wish to consider the added value in the AIX product before
you assume that it will be inferior, either technically or in the
marketplace.  As has been the case with AT&T's UNIX,  the software
product may have a bigger impact than will the machines it's
meant to run on.

(I speak only for myself.)
Chuck Karish	ARPA:	karish at denali.stanford.edu
		BITNET:	karish%denali at forsythe.stanford.edu
		UUCP:	{decvax,hplabs!hpda}!mindcrf!karish
		USPS:	1825 California St. #5   Mountain View, CA 94041