Mark -

On 9/14/06, Mark Ranft <[log in to unmask]> wrote:
> Pete,
>
> I for one do not think MPE is dead.  I have hopes.  I am an optimist.  But
> like so many others, I have heard over and over again how MPE is dead.  And
> now I am hearing how OpenMPE is dead.  And I am getting worn down.
>
> Are these people saying your ideas won't work, promoting the future of MPE?
> Why are these cynics part of this group?  What are they hanging out here
> for?  And what about OpenMPE?  Is it simply an organization set up as a
> buffer as HP eases MPE out of the picture?  What is OpenMPE really doing
> anyway?  Have they made any progress?

When I first came across OpenMPE recently, it was by accident.  It
looked like an organization to pick up the MPE ball that HP dropped,
or was trying to drop, and run with it.  I'm hoping that is still the
purpose, but starting to have my doubts.

> If there is a small role, I can play in the future of MPE, sign me up.  I
> will help.

I guess the first thing is to establish what the market is, and scope
of the problem:
1)  How many systems are still on the HP3000 / MPE, and would like to
remain there?
1a)  What is the yearly dollar amount they would spend to stay there?
1b)  What other resources would they be willing to commit (testing,
documenting, technical, hosting, etc)?
1c)  How long can they wait, or are willing to wait (which ever comes first)?
2)  A list of intrinsics that are needed to run their application set or system.
3)  What integration is needed to other systems?
4)  What client interfaces?

If there is enough market demand, are there enough technical resources
available to make it happen in a timely manner?

Next, have a plan to take MPE forward to make it play well in today's
world without sacrificing the simple and robust application
development platform with simple system management of application
resources, that focuses on one thing:  Solve business informational
problems as quickly, easily, and robustly as possible.  The hard or
complex stuff should be hidden behind the intrinsic wall and done
using Linux tools, and not exposed to the application development and
end user environment.

MPE may not be the long term answer, but it had the fundamental idea
before the inclusion and exposure of Posix years ago, and could easily
be morphed into a complete solution using a client interface that
would allow client apps like OpenOffice to seamlessly connect to
enter, display, summarize, search, and analyze the data stored within.
 The system manager "sees" a simple and straight forward platform and
resources, and the system operator "sees" a simple view of system
resources being used, jobs running, users logged in and their
processes.  Yet, all the power of Linux sits purring underneath,
easily exposed and/or controlled by a utility app using a hook in the
PM intrinsic interface (all PM intrinsics or utilities would be
blessed by OpenMPE, or the site would be on its own).  All application
programs are restricted to the intrinsic interface minimizing the
chances for bugs, and making debugging far easier.

It *appears* to have been done commercially with MPUX.  As MS squashed
all commercial competition that got between it and money until open
source Linux came along, the only hope of long term success, I believe
lies in making OpenMPE open source, but friendly to 3rd party closed
source.

My $.02
- Pete