HP stuff (was HP's not cool)

From: Eric Smith <eric_at_brouhaha.com>
Date: Tue Nov 23 00:42:05 1999

"Jay West" <west_at_tseinc.com> wrote:
> Yes, you do have to remove memory cards from a PDP-8 to add I/O cards, *IF*
> your backplane is full.

No, you don't. You've obviously never used a PDP-8. Maybe you're talking
about an 8/e, 8/f, 8/m, or 8/a.

> Of course, with the -8 you can always add another internal
> backplane, but that's hardly an elegant solution.

No, that's not true either. There's a finite limit to the number of
backplanes you can use. But that only applies to Omnibus 8s as well.

> I know both instruction sets well (albeit HP's much better). I don't find
> the 21MX instructions to be any more or less like the -8 than most any other
> machine out there. I suspect that the 21MX set is similar not because of
> the -8 being a "forefather", but because most instruction sets have a large
> number of similarities on any set of machines (except Pick virtual
> assembler, that is truely a unique animal). I just can't agree that the 21MX
> instruction set was derived from the 12 bit DEC stuff. There's nothing
> unique about the DEC set that is in the HP set.

Very few machines other than the DEC machines and those inspired by them
use the same sort of memory addressing scheme with small pages, a bit in
the instruction to select between a global page and the current page, and
a bit for indirection. The indirect bit appeared by itself in earlier
machines, but usually they had a much different mechanism for maximizing
the utility of the addressing bits in the instruction word. HP and DG
copied it, along with much else from the DEC 12 and 18 bit machines.
Similarly with the "microprogrammed" operate instructions. AFAIK, DEC
was the first vendor to use that concept in a mass-produced electronic
computer, but they were copied by several vendors.

> Ok, I have to challenge you here. Other than basic instruction set
> similarity, do you have any historical evidence that the 21xx instruction
> set was derived from the DEC machines?

Only the anecdotal evidence that HP used many PDP-8s in house before they
decided to enter the computer business themselves.

But really, are you trying to claim that the HP designers were NOT
familiar with the PDP-8 family, and designed something that purely by
coincidence happened to be very architecturally similar? That seems
like an incredibly unlikely scenario. Even HP doesn't generally have
such a terrible case of NIH syndrome as to completely ignore what the
competition is doing.
Received on Tue Nov 23 1999 - 00:42:05 GMT