DMF-32 (and clones) versus DH-11 (and clones) versus DZ/KMC-11

[Henry Spencer]

Some comments on Doug Gwyn's comments:

> The KMC11 is not at all horrid to program, if you use the software
> supplied with every AT&T UNIX release since System III.  ...

Quite right.  What I should have said was, roughly, "the KMC11 is horrid
to program, unless you have the AT&T software package, and AT&T has
packaged it properly and it's all there and working, and it is really
suitable for your particular needs".  The reservations in that don't
indicate bad experiences, just paranoia.

> There should be no "Unibus problems" with the KMC11B.  It can and
> does serve as a bus master, but so what?  That is part of the Unibus
> design.  ...

Of course having other bus masters works fine... *if* the things are
designed right.  Some aren't.  People I consider reliable have mumbled
about things like bus hangs and having to make hardware mods to KMCs
to get them to run reliably.  Perhaps these problems have been fixed,
or were local to odd Unibuses like the 780's UBA.

[It is a mistake to assume that all Unibuses are alike; they aren't.
The Unibus specs permit a fair range of behavior patterns, and the only
way to design boards that work on *all* Unibuses is to work entirely
from the spec.  Not all designers are conscientious enough to do this.
Not even all Dec designers.]
-- 
				Henry Spencer @ U of Toronto Zoology
				{allegra,ihnp4,linus,decvax}!utzoo!henry