Some time ago, just immediately after HP's announcement of its end of sales
and support of MPE, I wrote here that MPE might well have a brighter future
away from HP than it did under its stewardship. Indeed, I've long believed
that HP has been a very poor steward of MPE, not realizing the quality of the
operating system it had.

Although I was perhaps too subtle, I've also long argued prior to November
14th that the best course for CSY was to form itself into an independent
agency, perhaps a wholly-owned subsidiary of HP, even going so far as to
propose a logo for the new company:

     http://aics-research.com/csylogo.html

It's also been evident for quite some time to anyone who cared to look that
the business model that CSY was working under was deeply flawed and virtually
guaranteed to fail, resorting to charging exhorbitant prices for identical
hardware to an increasingly smaller installed market. At the inevitable
endpoint of this highly predictable failure, it is the nature of a middle
manager to say, "Well, geez, folks, we gave it our very best, but we just
couldn't make it work," and then move on to other, hopefully better
assignments.

However, this is not the attitude that would have ever been adopted by any
general manager/president of an independent company, if CSY had become that
sort of agency. The problems associated with CSY's economic model would have
been immediately apparent and they would have been solved. There's nothing
that so focusses the mind as the prospect of being hanged in a fortnight, and
every small company president operates constantly under those rules.
Entrepreneurship and imagination come relatively quickly even to the most
dull among us under these conditions. Given the quality of product that MPE
is, it almost certainly would have greatly prospered if had operated under
the same levels of entreprenurial enthusiasm that you characteristically see
in an Adager or a Robelle.

HP's current plans are simply to bury MPE. After November 1, 2006, HP will
not support or sell any aspect of MPE. At that point, they will no longer
discuss MPE any further.

Given those plans, I have been speaking for the last several months with
Susman & Watkins, a corporate law firm in Chicago, about the users filing
suit to legally obtain the right to use MPE, unfettered by any further
interference from HP. The theory, untested at this juncture, is that HP's
plans to discontinue MPE are much like you putting your garbage on the
street. Once you've abandoned it, it's free to be picked over by anyone who
passes by. But even more than that, the users have paid many times over for
the development of MPE and have an obvious intrinsic right of ownership, in
and of themselves.

While the ideal condition would have been for the separation from HP to be
wholly civil, having HP actively help in the construction of an emulator,
providing complete access to the source code, it has become increasingly
obvious that they're not going to do that. Instead, their current plans are
the most cynical and mean-spirited process possible. If an HP-sanctioned
emulator is to go forward, a user will only be able to use a licensed version
of MPE on an emulated system if he destroys an equivalent HP3000 system. NO
AGENCY THAT PURPORTS TO SPEAK FOR THE USERS SHOULD AGREE TO THESE TERMS.

Not only will no user do this as a practical matter, it's clear that HP's
intention is to limit the MPE user community to the number of licenses at the
time of its death, thereafter drawing the population constantly downwards
over the years. This is completely opposite to the future that I believe is
possible, where MPE would be distributed world-wide, at prices comparable to
Linux distributions. There is still enormous value in MPE and IMAGE, and it
is entirely possible, if given the chance, that MPE could easily spread and
grow into a user community of hundreds of thousands. (Do not confuse the
exorbitant prices that HP was charging for the HP3000 that resulted in its
constant decline in popularity with the inherent value of MPE.)

The construction of an emulator and the free use of MPE are two separate
issues. There is more than sufficient information available in the public
domain to construct a PA-RISC emulator on inexpensive Intel hardware, most
likely using Linux as a base. But even more importantly, there really isn't
much that needs to be changed in MPE. It could work as it does right now for
many years to come. Almost all of the future evolution will lie in the
hardware, and a virtualized emulator's I/O into a Linux base will take care
of much of that change automatically.

There are also apparently legal issues with some of the MPE/iX POSIX shell,
but these can be dismissed with the insertion of a simple JMP op code into
MPE, so that at point where you type ":sh", the jump moves you to Linux
directly. There are many obvious advantages of doing this. Whatever legal
problems there are with HP's POSIX would be immediately disposed of with a
single jump command. But even more importantly, the constant day-late,
dollar-short attributes of porting routines such as samba, vi, etc. that have
plagued MPE/iX would be completely eliminated. If Linux has those routines,
so those this newer version of MPE.

Can you maintain and grow MPE code if you don't have access to the source?
Yes. In 1976, we extensively modified HP2645 terminals to become AICS
Research System 2000 Word Processing Terminals, doubling the amount of
internal code in the machines. Although we purchased the source code from HP
to do this, we could have done it even without seeing the source. At every
point where we wanted to modify a behavior, we simply put a JMP op code at
the beginning of HP's routine, essentially abandoning HP's code in place, and
wrote our own new code. Eventually, approx. 100 of these jumps were peppered
into HP's terminal code, completely changing the machine's external behavior.
The same can be done for MPE. If you wanted a new LISTF format, for example,
a jump could be put at the beginning of that section of code, allowing you to
write whatever now-fully documented code you feel is necessary.

HP, in conjunction with the OpenMPE organization, is planning Yet Another
Survey, this one larger than the last, in order to "more fully understand"
the wishes of the user community. I believe this to be nothing more than a
delaying tactic, to put off whatever announcing whatever decisions that HP
has made until the end of the upcoming HP World Conference, long after Carly
has spoken, in order to forestall any possibility of any sort of disruption.
This will be the last HP World Conference for virtually every HP3000 user,
and may be one of the last conferences in general. As a consequence, if you
have a point of view to be expressed at this conference, this will be the
place to make it. HP is going to kill a number of other operating systems in
the near future, and how they handle MPE will set the tone for the other user
groups as well, so people will be watching.

Personally, I've gotten to the point that I am no longer concerned about the
legal niceties and believe that servers should be rented on the Principality
of Sealand, far beyond the reach of HP Legal, to distribute a user-supported
version of MPE from there, even if all future development has to be done
covertly. (For more information on Sealand, please see:

     http://www.wired.com/wired/archive/8.07/haven.html
     http://www.usatoday.com/life/cyber/tech/cti162.htm )

I no longer have any faith that anyone at HP in a position of responsibility
is looking out for the best interests of the MPE user community. The people
in the former CSY are being personally graded by how well and how many MPE
users they move over to HP-UX, thus they are intrinsically working at
cross-purposes to the MPE user community. I believe it may be time for the
users to begin to manipulate the future of MPE for their own best interests,
with or without HP's consent.

Wirt Atmar

* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *