>Is that so, can I have a HP9000 with MPE?

Well yes, but they call that an HP3000.

>Some I performance would be
>better. HP3000 is a piece of hardware, but a HP3000 is not a HP9000.

Right. But many of the HP9000 and HP3000 systems are the same
hardware-wise.

> Some
>new fast buscontrollers are not available on HP3000, but are on HP9000
>available.

The N-, L-, and V- class servers aren't available as HP3000 systems.
Yet. My understanding is that HP is planning an HP3000 based on the
N-class chassis (the 999? I dunno).

Also, peripherals might be available for HP9000 before they are
available for HP3000 (for instance, DDS-3, DLT libraries, Autoraid
12H, Fibre Channel, Gigabit Ethernet, were or are all available for
HP9000s before HP3000s). This is because HP wrote drivers for HP-UX
before they wrote the drivers for the same hardware for MPE/iX. This
is because 1) there are more HP-UX systems out there, so thus more
demand for new peripherals, so HP makes more money faster by
supporting HP-UX first and 2) I think HP-UX drivers are easier to
write because "programming culture' among UNIX programmers seems a
little lax compared to the culture I believe exists for MPE/iX. As a
case in point, consider how many problems there have been over the
years with TCP/IP implementation on MPE/iX, and consider how that code
is really ported HP-UX code. Further, MPE/iX sometimes demands more of
I/O drivers than HP-UX does. I hope someone who has experience with
MPE/iX drivers can correct me if these perceptions are incorrect.

I want to point out, though, that the HP3000 959 came out *before* the
identical HP9000 K series systems, back in 93 or 94 I think ... about
the only time I can remember the 3000 getting a head start over the
9000 systems.
>
>The hardware does not match any more.

Yes. The HP9000 systems seem to be a generation ahead of the HP3000
hardware, for both technical and business reasons.