Ted asks:

> We are running a 979/400 with roughly 150-200 concurrent users.  Our
>  systems people (who by the way hate the HP3000 and love NT) say the life
>  expectancy of this machine will be 2-3 years and then have to be
>  replaced.   Since we are moving our financials off this machine (and
>  onto NT) I would think we would be able to keep this box running a lot
>  longer than that.  Anyone care to comment.

Nonsense.

If by lifetime, your friends mean hardware, I would say that the NT boxes are
much more likely to be short-term devices (2 to 3 years). The 979/400 will
serve you well for a great deal longer than that, if you wish. Most of our
customers have traditionally run whatever version of the HP3000 they have for
somewhere between 5 to 10 years before they move up, unless of course, they
find themselves in a rapid expansion phase of the business, but that's a
circumstance to be celebrated, not complained about.

However, if they mean software lifetime, they're obviously showing the
brashness (and occasional ignorance) of youth. For a simple example of the
lifetime of MPE-code, please see:

        http://aics-research.com/series33.html

There are three HP3000s in the photo, our first Series 33, purchased in 1979,
a 922, and Micro XE. Only the Micro is still run on a regular basis, but the
moral is the same, whether all of these machines were on-line or not: We
began writing code for the HP3000 in late 1976, borrowing an HP3000 at night
from our local HP office. That code was moved to our Series 33, once it was
installed. Virtually of all that code is  still in operation today, moved
seamlessly and effortlessly and flawlessly from machine to machine,
generation to generation, as we've purchased them.

Return to this question in 25 years and ask your NT friends if they can say
the same in the year 2025 for their NT boxes. I'm sure that I'll be able to
say the same for the HP3000 then -- with the obvious exception that the code
we first wrote in 1976 will be 50 years old by then.

Wirt Atmar