According to Data, Not Dead
OpenMike is a guest editorial space and talking point for the 3000 community to express its views. In an era where emotion and analysis run abreast for the communitys attention, we want OpenMike to be a forum for the way you feel about the future and what you believe is important to the 3000 community. Send your contributions of less than 1,500 words for consideration to firstname.lastname@example.org.
By Gavin Scott
The problem with comparing the demise of MPE with the death of an individual is that in the case of MPE theres really no need to be in a hurry to start digging a grave.
Far from being dead, the HP 3000 has probably just entered into a period of steady decline which can be characterized by a half-life just as in the case of the decay of a radioactive isotope. My rough guess for the half-life of MPE at this point is probably around 2 years. The standard wisdom seems to be that you need 10 half-lives before theres nothing significant left of the original sample, so from this we would expect that 20 years will elapse before the last 3000s pass out of use.
From this metric one might predict that in 2007 there will still be one-fourth the number of 3000s in use in production as there are today, even though HP will have stopped providing any support for them at the end of 2006.
Lets look at some data from the most recent Interex e3000 Migration Survey...
52 percent of customers estimate that the migration will take more than two years. So theres our first two year half-life.
When asked what their migration plans were, 23 percent said that they planned to continue on MPE past 2006, so theres our one-fourth of all systems still around in 2007. We have almost 25 percent planning to stay on MPE, and I suspect that many peoples estimates of how quickly they will have migrated off of the 3000 are quite optimistic considering most people havent started the process yet.
In the same question about migration plans, 35 percent said they would either migrate their current application code or rewrite their application on a new platform. Both of these are options which may experience a relatively high chance of cost- and time-overruns, and even outright project failures, extending the time that these people will be on MPE. And even the 34 percent who plan to simply buy a new application on a new platform and the 7 percent who plan to simply stop using their MPE applications by 2006 will have a lot of work to do to accomplish this in a reasonable amount of time.
And 58 percent of customers expect to (or will be forced to?) complete their transition using only their own internal staff, which will probably result in a more leisurely pace for the process.
Based on the amount of money Ive seen made over the years as a result of renting 3000 software to people who were going to be off the platform in six months and who were still there five years later, I think the user base is still going to be quite crowded in 2007.
Also not all customers have yet experienced life out there in the big, wide, non-MPE world. Myself I find using Unix and Windows systems quite interesting, and rather like visiting some exotic foreign country. The scenery and architecture is quite interesting, though eventually you find your self wishing to be home in your own bed and thinking why cant these people just do things the way we did back home?
The average emigrant from MPE would be well advised to come prepared with a complete compliment of remedies for motion-sickness, indigestion, and stress.
And some of those who depart are going to find after some time that they really would rather go back home, even if home means living on an unsupported platform.
So... Obsolete? Perhaps. Dying? Maybe. Dead? Not for quite a long time.
Gavin Scott is vice president of Allegro Consultants, a support provider and product developer of solutions for HP PA-RISC systems, including the HP 3000.
Copyright The 3000 NewsWire. All rights reserved.