What a coincidence...

Just a few days ago, I talked to a customer who is switching his
HP 3000 to "archive lookups only" mode because they have migrated
to an NT & Oracle based package (or have been "migrated" by their
application vendor?).

He was not quite happy with the new "fancy schmancy" solution.

The NT database server has 1 GB of main memory, much more than
his 3000 had, but still doesn't "fly". The database needs constant
fiddling to keep it alive, whereas on the 3000 he had some batch
job running once per week to auto-increase dataset capacities etc.

His 3000 had downtime problems maybe once or twice a year, whereas
his new collection of application servers (notice that the "fancy
schmancy" architecture needs more than one single computer) tends
to have something like one app server outage per week.

He also had some comments on spooling or spoolfile handling on NT,
but I don't recall the details (something like "when a printer goes
down, I cannot simply SPOOLF x;ALTER;DEV=another").

As far as I got the impression, he'd rather be back on the 3000.

Lars.

(not meaning to say that NT or Oracle or applications based on those
technologies are inferior; it's just that this customer's experience
was "need more hardware, more software, more manpower and still have
to struggle to get the job done" whereas the 3000 solution was more
like a fridge... turn it on and let it work for you).