Paul, responding to Zelik, writes:

> Some companies, including mine, try to take the less risky approach.
> We try to select vendors that we consider safe.  If that vendor goes
> out of business or their product becomes unsatisfactory, then we start
> finding a new one or at least have a plan if things fail.  Compare this
> to your desktop or laptop.  After so many years I would assume you
> start considering replacing it.  You know you need to do it but
> oftentimes you procrastinate.  Eventually, you replace it because you
> know that some day that hard drive or screen or whatever will fail or
> it just isn't fast enough or capable enough.

For a mission critical system you design a modular solution to limit
downtime and ease change.  "After so many years" we expect to upgrade a
component (like the processor or even an entire box).  We do not, however,
expect to replace 75%+ of the hardware and 80%+ of the software.

We didn't plan for the system being discontinued because HP never told us
that it would be or what the numbers really were so that we could figure it
out on our own, and their prior business model suggested to us that this was
not an issue.

None of us started down the MPE path thinking that HP was not a "safe"
company with which to do business.  They were (and if you count Agilent,
still are) a showcase of partner focused business entities working together
(maybe not as well with certain changes in the past 18 months, but...).

And the concepts of the "vendor going out of business" or their product
"becoming unsatisfactory" don't apply to our situations because the vendor
is still in business and the product is still great.  The problem is that
the vendor has decided that their profit potential is greater on a different
set of their products than on the set over which we built our relationship
and they are *forcing* us to switch to remain partnered with them.

I do think you bring up a good point by making a comparison with desktops.
Filtering through the marketing crap, the message I'm getting from HP isn't
really all that off the wall: the world of computing has for some time been
maturing towards a low profit margin, interchangeable component oriented,
commodity environment.  (There's a description <smirk>.)  HP feels that
investing funds in relatively expensive proprietary OS architectures is
wasteful; and will not get them where they want to go.  (Don't anyone tell
IBM.)

Fine, no problem.  To a great degree, I agree with that point.  My issue is
that the ideas I'm seeing presented about what the future will be do not
change my understanding of how best to run my business today and in the
midterm future.  The commodity model lends itself to distributed processing,
distributed storage and distributed control, not the all in one package we
have with the HP3000.  A fully distributed system (think farm of HP9000s
each with its own application subset) is only ONE network arrangement and
for some type of businesses/applications it is not the best model.

I have never forgotten a lesson I learned in a Strategic Management class in
college a few millennia ago: for each type of business there is a model that
works best.  Specifically, if the work being performed is very repetitive,
and requires very little innovation on the front lines, a beaurocracy is the
best model.  Situational exceptions are passed upward until a level is
reached where the problem can be addressed or booted out of the entire
system; but 98+ percent of the work is performed using fully documented
mundane procedures.  I hate the frustrations I feel trying to get anything
from a government office too, but I recognize why the organizations are
setup the way the are.  Similarly, if you are doing consulting work, your
organization must be very flat and there must be a greater amount of control
given to the front line staff to innovate or they cannot accomplish the
business goals of meeting very varied customer needs.

Computer systems are not so different: for each usage type, there are models
which will work better than others and the choice should be made by the
application/business need, as your comment suggests.

>  If you can't find a
> replacement for software on your HP3000, maybe you need to re-evaluate
> your needs.

But I think that also misses the point of our migration concerns and the
OpenMPE push: there are many businesses for which the HP3000 and the
applications that have been designed to run on it are *exactly* what the
firm needs for now and the foreseeable future.

And how is revisiting an analysis of what the company does and needs, going
to help anyone see that differently?  If we determine that the HP3000
running application suite X is the best solution for our business needs, the
only way that will change is if we look again and say "oh but that option is
no longer available".  It may not be available, but it is *still* the best
option.  Can we find a replacement?  Yes.  Is it a better option?  No,
otherwise we would have already chosen it instead.  When the answer is still
going to be a custom application running on a single centralized server with
a tightly integrated DBMS, now we have to look for another system capable of
handling it and then we're looking at other vendors because HP is making us
leave their version of that answer.

MPE's strength as a tremendous OLTP server with stability, security and
(within reason) ease of application development and maintenance, is *not*
something we can replace with a wall of WINTEL servers running an off the
shelf application if we want to keep our companies alive.  There is not an
application available to meet our needs other than our own custom design.
Can we completely restructure our business to use canned applications for
little pieces of our business, and hope that the synergy of the nine systems
exceeds the one system we have now....sure.  Does anyone else think it's an
excellent opportunity to get to talk to bankruptcy lawyers....um, yeah.

> The warning signs of HP's announcement to terminate the HP3000 have
> been there for years.

Yeah, the announcements by the number 2 person in HP and the number 1 person
in CSY for the past three years at HP World that MPE was strategic and was
in HP's future, etc..., well, that was just their own "special" way of
saying "Get off this platform".

>  Am I happy that HP has decided this?  Of course
> not.  But none of us should have been surprised by this.  Now each of
> us have to make the decision to continue on a risky path or move onto a
> more stable path.

Raise your hand if you think HPUX/Oracle or WINTEL/SQL systems are going to
be more "stable" than MPE/Image.

>  Or are there ways to make the risky path less risky?
>  It sounds like that's what we're trying to do here.

Yes, but don't muddy the waters by suggesting that better analysis and
designs of our own systems would have mitigated this problem.  HP pulled the
rug out from under us, period.  IBM has basically the same model and is
making good money with it.  HP used a crappy marketing mindset with regard
to the HP3000 and seemingly never wanted to invest in it once it had opened
the door for HP to enter the midrange server market.  Now HP wants to get in
bed with Microsoft and Oracle, and if that were the best solution for any of
us we would already be running it on the mission critical systems instead of
just on our intranets and reporting/marketing data warehouses.  If the AVP
can't get one of his custom data "views", it won't cost us customers who
walk away from us because they're having a bad day and not being able to do
a transaction on their account with us was their last straw.

> Criticizing
> another person's approach to software accomplishes nothing.

I disagree.  If the criticism is constructive it may *teach* a person
something.  :)

Rs~

Russ Smith
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
* * * * * * * * * * * *
The opinions expressed in this email are mine, and are not meant to reflect
those of any other party.  The subject matter herein is intended solely for
the named recipient(s) of this email.  Spellcheck cancelled.
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
* * * * * * * * * * * *