[comp.sys.apollo] HP, OSF, etc...

librarian@ADUS.ECN.UIOWA.EDU (ADUS librarian) (12/21/90)

This article was co-authored by Dave Shaw (the ADUS Librarian), and
Aubrey McAuley, editor, HP/Apollo Workstation Magazine......


In an effort to help clear up some of the misconceptions that have
appeared here recently, the following information is offered regarding
the seemingly changeable position that HP has taken regarding OSF and
the Apollo DN (and HP 9000) line.

   1. At the ADUS conference in early October, HP/APOLLO announced
      that OSF would not be supported on the DN line.  *MANY* of those
      present complained.
   2. HP sent a letter dated Nov 5 which stated that they had 
      reconsidered their plans for OSF, and were going to offer OSF/1
      on DN machines which were upgraded to the 68040 processor.  The
      reason given was that enough people complained. (Although it may
      be no consolation, it should be noted that HP does not plan at
      this point to support OSF on 68040-based 9000/300s at all.)
   3. The December WORKSTATION magazine, which went to press before
      the Nov 5 letter, printed the HP position presented at the
      conference (i.e., the outdated information.  At the time
      WORKSTATION went to press, HP was still saying it was
      considering its options. WORKSTATION regrets any confusion that
      may have resulted from the poor timing of events.)

To the best of our knowledge, HP plans to stick by the decision noted
in 2 (i.e., it is the most recent word out of Chelmsford). WORKSTATION
personnel pressed HP pretty hard on this one to make sure they were
really committed to the 68040 support, and it truly seems they are
sure about it. (In the interest of clarity, an article about these
topics that will appear in the January issue of WORKSTATION (soon to
reach your  mailbox) appears at the end of this posting.....)

------
In addition to the discussion regarding OSF, many other issues of
software/hardware support/longevity were discussed at the ADUS
conference. These issues involved HP hardware including the DN line,
the Series 400 line, and the PA-RISC line, as well as peripherals,
buses, etc. Virtually all HP-supplied software, from OS's to shells to
compilers was included in these discussions. 

On the last day of the conference, an extensive table was presented by
HP management which indicated the expected fate of each combination of
software/hardware. As an example, the future of Domain/OS on the DN 
line (except for the DN550/580) was said to have 'no planned
termination date'. As another example, OSF/UE and OSF/DCE was said to
be in the works for the DN line (except the DN550/580). This table
had been hastily drawn together, and a promise was made that an
HP-approved version of the table would be sent out to conference
attendees within a week (or two?). This table of information has yet
to be mailed, and the reason given has been that HP is still
considering the many decisions involved. This could be read as saying
that *your* desires/needs still can change the outcome. 

If you feel strongly about some aspect of future HP hardware/software 
support, VOICE YOUR OPINION! (and *not* just to comp.sys.apollo- not
everyone reads this stuff, including HP personnel). 
It might not be too late.....        

Happy Holidays....


WORKSTATION text follows----
******************************************************
HP TO SUPPORT OSF/1 68040 DNs, BUT NOT HP 300s

By Aubrey McAuley
Workstation Staff
        NEW YORK -- As the Open Software Foundation was demonstrating
the results of its two-and-a-half-year Unix development effort here at
Unix Expo, Apollo users got a reprieve from obsoletion with HP's
announcement that it will support the OSF operating system on a few 
older workstations after all.
        After a last-minute stamp of approval by high-level officials,
HP announced at the show that it will port the OSF/1 operating system
to Apollo DN-series workstations that have been upgraded to use
Motorola's 68040 processor. However, under "the current plan" HP  will
not support the OSF/1 kernel on similar 68040 upgrades to HP's 9000
Series 300 workstations, according to HP OSF project marketing manager
Bill Kelly.
        Kelly said HP decided to support OSF/1 on upgraded DN-series
workstations after listening to numerous Apollo customers who said the
lack of OSF support would force them to "throw away" their current
Apollo systems without fully depreciating them. But Kelly said HP did
not receive such a strong negative response from Series 300
workstation users, who remained mostly indifferent throughout the
controversy which erupted when HP announced at the Apollo Domain User
Society (ADUS) conference in early October it would not support OSF on
older Motorola-based workstations.
        "There has not been the same kind of groundswell from the HP
community, and I think the issue there is that people with Domain knew
they had to move to OSF," Kelly said. People with HP-UX don't feel the
same pressure. In the Domain world, people have wanted to go to a
'real Unix' open environment for a long time, but they weren't willing
to give up Domain functionality, which I think we've covered with the
DCE. HP-UX has already been looked at as a standard Unix 
implementation."
        The move means Apollo DN 3500 and 4500 users who purchase a
board upgrade to the 25-MHz 68040 system, which has informally been
called the DN 5500, will receive the same support for OSF/1 and later
releases that 9000 Series 400 users will get. In contrast, HP 9000
Model 375 users who purchase a board upgrade to the 68040, which has
informally been called the Model 380, will not be able to run the OSF
operating system kernel.
        All HP-UX systems, however, already support OSF/Motif and will
eventually support the OSF's Distributed Computing Environment, two
major components of OSF/1. HP Series 300 users who want to run the OSF
kernel will be forced to purchase new Series 400 machines or HP's
upcoming RISC boxes, a choice HP expects many users to take, Kelly
said. But current HP-UX users also feel less pressure about moving to
OSF because HP-UX, unlike Domain, will be supported on the  new RISC
machines, he said.
        Kelly said HP plans to follow the initial HP_OSF/1 release,
set for the second half of 1991 on the new RISC boxes, with a
maintenance release approximately six months later that will be
C2-certified and will include support for Apollo 68040 upgraded
systems. The OSF/1 version supported on the 68040 DN series
workstations will only run in a 2-D X Window graphics mode and will 
not support Apollo G*R or PHIGS graphics, although all Apollo graphics
displays except some older 4-plane color systems will be supported in
this mode.
        HP says it has struck a good compromise in the new plan
between the demands of the installed base and the realities of an
increasingly competitive workstation market. However, HP 9000 Series
800 and Apollo DN 10000 workstation users will still be left without 
OSF/1 support under the deal unless they switch to HP's newer
generation of RISC machines. 
        Yet HP has also pledged to support the OSF operating system on
Series 800 and 600 RISC multiuser systems and servers, which are
binary compatible with the 9000/800 workstation models but do not
require support for workstation-style graphics such as the CHX, SRX
and TurboSRX. Although HP has said its new RISC machines will include
desktop, deskside and superminicomputer systems, the availability of
OSF on its high-end multiuser systems-- set for sometime in 1992 -- is
seen as vital to HP's continued success with Unix in data processing
environments.
        "It's really a question of when we feel like there's enough
functionality there to release it into the commercial marketplace. In
the workstation market, we know that customers are a little bit more
gracious about technology. They need it more, so they accept it more 
readily," Kelly said.
        "Certainly by the time we roll around to the next Interex
meeting [in August of 1991], we'll have all that nailed down," Kelly
said. "The volume of the Series 800 has never really been significant
in the workstation space. When you look at it from an installed base 
perspective, there's just not the same kind of requirement as there
was in the Domain line. We think that people will be moving so quickly
to the new PA-RISC workstations because of the price/performance
attributes of the product that the 800 will quickly become
[unattractive]. On the 800 multiuser line it's a whole different
ballgame" because the new PA-RISC line will not initially include
multiuser systems, he said.

Aubrey McAuley- wks@pcinews.lonestar.org
******************************************************

asherman@dino.cpe.ulowell.edu (Aaron Sherman) (12/29/90)

>>>>> On 21 Dec 90 05:42:51 GMT, librarian@ADUS.ECN.UIOWA.EDU (ADUS librarian) said:

     ADUS> WORKSTATION text follows----
     ADUS> ******************************************************
     ADUS> HP TO SUPPORT OSF/1 68040 DNs, BUT NOT HP 300s

     ADUS> Expo, Apollo users got a reprieve from obsoletion with HP's

I find this kind of attitude to be sickening. Why don't people wake up
and realize that HP's just kicking us in the teeth and then saying "Aren't
you glad I let you live". They're mocking us by taking away one of the best
development platforms ever to have been designed, and replacing with
a brand new (read buggy as hell) Unix clone. What's out consolation?
We're not "obsolete". Bah! The system managers here have found a solution.
There's a third party company that we're getting Apollo equipment support
from now, and we'll just never "upgrade" to OSF when it come along.


HP: You're losing your customers... It's time to think about what you're
    doing!


			-AJS

---
| asherman@dino.cpe.ulowell.edu	or	asherman%cpe@swan.ulowell.edu    or  |
|                  {backbone}!ulowell!ul-cpe!asherman                        |
| "The protection of private data.... can therefore be circumvented.... but  |
|  this, of course, is cheating."  -- 'C++ P.L.' pp 137 by Bjarne Stroustrup |
------------------------------------------------------------------------------
--
| asherman@dino.cpe.ulowell.edu	or	asherman%cpe@swan.ulowell.edu    or  |
|                  {backbone}!ulowell!ul-cpe!asherman                        |
| "The protection of private data.... can therefore be circumvented.... but  |
|  this, of course, is cheating."  -- 'C++ P.L.' pp 137 by Bjarne Stroustrup |
------------------------------------------------------------------------------