Jeff writes after Nick:

>Alright, if everyone else is going to participate, here's my $0.02:

I guess I gotta as well! In my case, this story is a "too successful"
story.

>> My introduction to the HP3000 came in 1977 with the Series II (I knew
>> I could gray-hair you guys).

>Mine was in 1976, but at the ripe old age of 17.  Had HP2000 in 1975.

My first HP experience was with an HP2000 in 1970. My first HP3000
experience was in 1975.

My "too successful" story takes place in 1982 or so. I was working for
abacus systems at the time. They had a customer involved in nuclear
engineering that had an engineering program than ran on the
supercomputer of the time (I think it was CDC-Cyber - but I'm getting
foggy in my old age). The program reduced and reported data collected
from nuclear power plants and took some number of hours to run. Each run
cost something on the order of $20,000 which was charged back to the
engineering company's customer (with a suitable uplift to add to the
bottom line).

The timeliness of the data wasn't important - whether or not it was
returned in two hours or two days was immaterial. The company was
looking into bringing in house their computer processing and we were
bidding on the project. Of course, we were bidding the venerable HP3000.

I took the engineering program (written in fortran) and converted it to
SPL and used every trick in the book that I could think of to make it
perform. I managed to get it to reduce and report the same data but in
about 24 hours. We figured that we the customer could replace the CDC
and pay for the 3000 with a couple of runs of the program. Of course, we
won the initial bid.

However, as luck would have it, some muckety muck at corporate realized
that they would be unable to justify the same cost to their customer for
the machine time. Something about contracts and government audits and
whatever. Since they were adding a percentage to the actual costs, they
reasoned that if the cost of the computer time went down, their take
would also go down. So, the project was dropped.

Too bad. I was proud of the fact that that little 3000 could compete
with the supercomputer of the day in terms of bottom line costs. If the
prospect weren't in the service business and padding their outside
costs, they could've reduced their machine costs by orders of magnitude
- paid for the machine within a week and been happy ever after! Oh,
well!

:-)

Regards,


M.