I think the problem in having benchmark figures for HP3000's is in the heads of
the decision makers.
They never ask: how many users can run in parallell having this amount of
memory
and so on ... but they always ask for benchmark figures to make some
calculations
like: I get x TPC for y $ on Unix, this cost me z $ per TPC. Now I compare this
figure by NT and MPE ... ah, not available, so out ... and finally they compare
Unix vs. NT $s per TPC ... and MPE is out.
I know that's the wrong way to do it ... but who will tell this the "big blue"
damaged Managers ? Because most of them came from the MainFrame Area ... what
means IBM ... now they want to have client/server applications, and this sound
like Unix or NT but unfortunately never MPE ... and they do not know that they
are
wrong ...

Only a small remark, and probably I'm wrong ... but I take the risk to look
like a
fool ;-)
Best regards, Andreas Schmidt, CSC, Germany