I'd like to throw in my tuppence worth on some of this thread...
Ken Sletten B894 C312 x62525 wrote:

> It could perhaps be argued that making the following
> proposal should wait until IPROF is over;  i.e.:  maybe
> HP will overwhelm us with the massive scope and
> number of new enhancements to the 3000 that they
> are going to deliver in the near future, so that there just
> won't be anything else we need to ask for in the next
> five years+.  But given the current funding model for
> CSY and Roseville R&D, I'm willing to bet that's not
> going to happen.

I concur 100%.  The introduction of HFS, Posix, and Image/SQL
marked the peak of CSY development.  It has declined since.

> Since Proposition 3000 first exploded in the collective
> consciousness at IPROF-96, there have been many
> detailed suggestions for improving the 3000, from both
> inside and outside of HP.

The crux of Prop 3000 to me (as a primary instigator) was twofold:
* a perceived hardware death (no committment to even PA-8000 at the
  time) which signified the end regardless of CSY software actions,
* the irony of having HFS/Posix/bundled ARPA services to make us an
  open system, while not speaking (internally) with GSY for even the
  possibility of cross platform development (which continues to this
  day even with the addition of NT/Netserver group).

> (2)    Jeff Kell, Alfredo Rego, Allan Gardner, and
> Tony Furnivall, for being behind the genesis of the
> "Proposition 3000" initiative.  There was worry when
> it came out that it would be perceived as a very
> negative thing+.  Then a lot of HP engineers ended
> up handing out Prop 3000 buttons at HP World.  Plus
> Jeff for a lot of additional technical and strategic detail
> since then.

I don't mind repeating my sermons as long as the congregation
isn't entirely empty.  There is more than adequate growth in
GSY and HP-UX solutions developed across the 700/800 platforms.
This same software isn't out of reach of the 3000/900 platform
with some nominal porting effort; it's not reinventing the wheel
each time which they seem to keep doing.  It's a subtle point to
comprehend for some, realizing that we are 90+% of the way to
being able to compile any UX application in the posix environment.

> (3)    Duane Percox, for starting and carrying on much
> thoughtful discussion in this area since IPROF-96:
> Corba, funding models, compilers, Roseville, etc.

Roseville has been re-incorporated into the R&D model.  There is
not yet enough evidence to see if this will produce results, but
they are back "in the loop" and not just a black hole.

> Our current HP software support contract shows that
> the list support price for most of the add-on software
> products on our 3000 is $52 per month = $624 per
> year. We have quite a few products for which we pay
> this $624 a year.

You can pay that for Basic/V today.  When was your last
Basic/V upgrade?  "Support" involves answering questions
and isolating bugs (tech support) but should also involve
product evolution.  The latter is seriously lacking.

Furthermore, not everything is "$52/month" either.  Image/SQL
is quite costly.  We already put serious bucks into support
fees for that product.  So why are we still waiting on 32-bit
ODBC?  We got VPLUS enhancements, we got bundled ARPA, we have
had numerous significant enhancements everywhere *except* for
Image/SQL.  We voted, passed, and paid our extra support dollars
but still wait on 32-bit Image/SQL.  The "plain" Image for which
you do pay $52/month has given us DDX, dynamic rollback, and soon
to add B-tree masters.  Those aren't SQL/ODBC issues.  We got
slow and buggy drivers and still wait on 32-bit.  This has done
more than anything to shatter my confidence in any corraborative
evidence that funding dollars do anything to affect end results.

> SIDEBAR(s):  It would be a good idea to use some of
> that money to give bonuses to the CSY and Roseville
> engineers who are still working on the 3000, so that they
> have added incentive to keep doing that.  And it would
> be smart to use some more to offer significant bonuses to
> the well-known list of key HP engineers who have recently
> gone on to other things, to get them to come back to the
> 3000+.

The most significant contributions to the 3000 growth, particularly
in the posix arena have been Steve Elmer (initial ports, libbsd,
etc), Mike Belshe (NCSA httpd), Nicole Saeger (Open Market port),
Eero Laurila (:showconn/:abortconn and NM TCP/IP stack), Jeff Vance
(:listfile enhancements, CI enhancements, numerous others), and
other engineers doing almost "volunteer" work not officially
sanctioned (Mike Yawn has Java, but with half a blessing from CSY;
and Lars Appel for Samba, but his appears mostly volunteer as well).
Most of these folks are gone (the first four anyway) from CSY.
Then we have the customer contributions (Mark Klein for Gnu, for one)
also not CSY affiliated.

Neil Harvey <[log in to unmask]> writes:

> I'd also like to think that it could made voluntary, but then it would
> have to be VERY well motivated by H-P themselves.
> And I agree it MUST be used for R&D. Give some to Lars :)
>
> If a loner (I mean this in the positive sense) like Lars with help from
> the list can implement Samba, and other port Web servers and Posix
> tools, it seems to be that it's commitment that's required more than
> cash.

This summarizes my last point above in a nutshell.  Many significant
products have been individual efforts, usually unofficially sanctioned.
It doesn't take an army of engineers, just a few knowledgeable ones that
can deal with the porting issues.

> Let's reverse all this talk of migrating from MPE to UNIX. Maybe people
> are disenchanted with UNIX, and that's why there is an astonishing
> acceptance of NT as a mission critical platform.

You have unix (almost) on MPE with posix.  But fortunately, you aren't
stuck in the twilight zone.  If you get frustrated, just 'exit' and you
are back to a colon prompt.  We can't ignore unix applications, that is
what is driving the market right now.  But we can ignore the necessity
of
a unix hardware platform if we just "finish" and/or augment posix.

Food for thought...

Jeff Kell <[log in to unmask]>