Path: utzoo!attcan!uunet!steinmetz!davidsen
From: davidsen@steinmetz.ge.com (William E. Davidsen Jr)
Newsgroups: comp.unix.wizards
Subject: Re: O'pain Software Foundation: (2) Why is it better than AT&T?
Keywords: OSF, AT&T, standards, competition
Message-ID: <11006@steinmetz.ge.com>
Date: 26 May 88 17:32:41 GMT
References: <24369@pyramid.pyramid.com> <10978@steinmetz.ge.com> <14181@tut.cis.ohio-state.edu> <5085@nsc.nsc.com>
Reply-To: davidsen@crdos1.UUCP (bill davidsen)
Organization: General Electric CRD, Schenectady, NY
Lines: 107

In article <5085@nsc.nsc.com> glennw@nsc.UUCP (Glenn Weinberg) writes:

| 2) ABI's
| 
| 	Now, you say, "well, why doesn't everyone just sign up for an ABI,
| 	then?"  The answer is simple: because AT&T wouldn't let them.

  An ABI is not for a vendor, it's for a CPU type. There will be one for
SPARC, 386, 68020, and hopefully for VAXen. These are all markets in
which there are either multiple hardware or software vendors. ABI
describes how things call the O/S, and you don't need to "get one," you
just buy the standard and write to it.

| 	Also, keep in mind why the Hamilton group got together in the
| 	first place.  Initially, THERE WAS TO BE ONLY ONE ABI--THE
| 	SPARC ABI.  AT&T had to retreat from that position when they
| 	saw the reaction, and opened up the ABI idea to other vendors.

  I hate to say this, but is there some reason to believe that anyone
ASKED for more than one ABI?

| 	Of course, in the process they eliminated ABI's greatest
| 	attraction, that of being able to provide shrink-wrapped software
| 	for Unix.  (Now, AT&T and probably Sun will deny that a SPARC-only
| 	ABI was ever their intention, but if you believe that, I too have
| 	a bridge to sell you.)

  Beats hell out of me how this eliminates shrink wrap compatibility...
on the 386 market I'll be able to sell software for Xenix, V/386, IN/ix,
and probably some others I've forgotten. Right now there are (at least)
two standards, and I believe three.

| 3) Vendor neutrality
| 
| 	Many people have commented on the fact that OSF will be dominated
| 	by IBM and DEC.  This shows a clear lack of understanding of
| 	how OSF will be structured.  OSF is an independent foundation,
| 	with its own board, president and technical director.  In many ways,

  I'm sorry, if you believe that any foundation is "independent" of the
people who pay the bills, I disagree with you. If they want to be truly
independent, the board should be TOTALLY made up of users and software
vendors who don't sell hardware (or at least CPU's, I'd hate to count
the Microsoft Mouse and stuff like peripherals).

| 5) Licensing
| 
| 	Anyone who has seen the changes in the licensing agreements for
| 	System V from SVR1 through SVR3 must understand one of the
| 	greatest motiviations for creating the OSF.  AT&T has arbitrarily
| 	changed the definition of Unix, how it can be distributed, to
| 	whom it can be distributed, what it costs, etc., etc.  Major
| 	portions of the system have been unbundled or dropped altogether
| 	(e.g., DWB, man pages--did you know that the SVR3 license does
| 	not allow you to ship man pages?)  What will they change next?

  And here I thought that the reason was so that I didn't pay for
something I don't need... BellTech is shipping a UNIX runtime, C
compiler, and manuals (unlimited license) for about $400. If they make
money at that rate the license fees can't be too onerous.

| 	Of course it's possible that the OSF will play the same games,
| 	but I don't think they will, and besides, the MEMBERSHIP has
| 	control over the OSF's licensing terms.  Only AT&T has control
| 	over AT&T's licensing terms.

  True. In either case the user's have nothing to say about it. If OSF
gets really obnoxious you can always get the nice cheap proprietary o/s
from the vendor. I don't disagree with you on this point, but as long as
the user's have no say, we're at the mercy of whoever, anyway.

| 
| 6) Standardization
| 
| 	Sorry, the AT&T/Sun effort doesn't result in standardization either.
| 	First, even if you grant portability over all the ABI's (which I
| 	don't), as discussed above, not all vendors will be allowed to
| 	have an ABI.  Second, there is already the question of POSIX.

  AT&T has said they will conform with POSIX. Just getting away from
having versions of stuff for SysV, Xenix, Ultrix, SunOS, BSD, Apollo,
etc^3, would make me happy. Anything which gives me another set of
ifdefs rots.

| 	Third, since AT&T has shown no propensity to accept input regarding
| 	the contents of future releases, each vendor has been forced to
| 	extend Unix in its own way.  This is standardization?  At least

  Isn't having AT&T get Sun and Microsoft together to make a common
product for all of them an indication that they're trying to integrate
the major flavors? Sure, it's driven by the market, but it's happening,
and the software vendors (and developers) and users will be better off
for it. A common user interface which runs initially on both X and NeWS?
Can you give up all the proprietary unterfaces? I sure can.

                                 * * *

  I'm not ridiculing your opinions, I just feel that you are looking at
this from a vendor standpoint, and I'm looking at is as a developer and
purchaser of software. We have Suns, VAXen, Xenix, V/AT, PC/ix, and
about seven other flavors of UNIX here. I would love to think that we
could all share the same versions of the same programs, *and I will give
up a little of the benefits of either SysV or BSD to do it!*
-- 
	bill davidsen		(wedu@ge-crd.arpa)
  {uunet | philabs | seismo}!steinmetz!crdos1!davidsen
"Stupidity, like virtue, is its own reward" -me