Burroughs Mini-D

[Al Kossow]

Yesterday was researching the origins of the processor architecture of the 1970s small systems.
Three-chip Fairchild PMOS CPUs show up a lot, documentation is confusing. The parts are collectable now because they look 'cool'.
So far, I haven't found anyone who will answer my questions on the subject.
They were known as the Mini-D or Burroughs Basic Data System (BDS) micro-processor
Most of what I know so far has been pushed to bitsavers. The chip stuff is somewhat confusingly under http://bitsavers.org/pdf/burroughs/military/D-machine to keep all the information in one place.

[Paul Kimpel]

There's a capitalization problem in the link above. The correct link is:

http://bitsavers.org/pdf/burroughs/Military/D_Machine/

I worked for Burroughs in the Paoli (Philadelphia) area during the early
'70s, and certainly heard about the D and Mini-D machines, but never
knew much about them. I believe the full D machine was used in the
B700/800/900, and knew that it was a very soft architecture, but that's
about it. I have reached out to some colleagues from that time and we'll
see what they have to say.

Paul

[Scott Lurndal]

In Pasadena we had the B774/B874 which were used as data communications
processor (DCP) front-ends for the medium systems mainframes. Aside
from once using MSNDL to compile a network definition, I don't know
much about the underlying architecture.

[Timothy McCaffrey]

I knew somebody who helped design the D-machine, unfortunately he died about 20 years ago.
(nice guy, it was a loss).

I believe it was originally built for a defense contract (or two) and was later adopted
by the B800/B900 (aka. Small System) group. I think some/all of the Small System
development was located in Boca Raton, FL at one time (at least, some other co-workers
came from there when they closed down SS development).

- Tim

[Timothy McCaffrey]

I forgot to mention, the B800/B900 claim to fame was that it changed its instruction
set depending on what programming language you used. So, there was a Fortran
instruction set, and a Cobol instruction set, etc.

- Tim

[Al Kossow]

On Thursday, June 1, 2023 at 3:22:12 PM UTC-7, Timothy McCaffrey wrote:

> I forgot to mention, the B800/B900 claim to fame was that it changed its instruction
> set depending on what programming language you used. So, there was a Fortran
> instruction set, and a Cobol instruction set, etc.
>
> - Tim

I have a very large collection of material on the B1000 series which had S-Languages for each language

The 700/800/80/90 used a small microprocessor which was much more limited and unfortunately there are almost no surviving
machines or doumentation for them beyond what I've uploaded to bitsavers recently.
I've also discovered a few of the remaining B80's in the US that had been saved were scrapped
in the past decade.
I tried emailing a few people involved in the development in Scotland recently and received no replies.

I'm becoming concerned that what may have been produced about these systems has been lost.

[Al Kossow]

On Thursday, June 1, 2023 at 4:25:06 PM UTC-7, Al Kossow wrote:

> I'm becoming concerned that what may have been produced about these systems has been lost.

I went back and looked at the implementations information I have for the 700/800/80/90 systems last night and
they do run an interpreted system language on top of a microcoded 16 bit processor. Even on the B1000 side with
the internal memos that I have there was apparently no detailed documents on how the systems actually worked
beyond the communicates and data structures which changed with each release.
I'm assuming if you had the need to know, you would just look at the source code and none of this information was
really ever released to customers.

The situation is much worse for the non-B1000 small systems. I've not even been able to find field service documentation
beyond some schematic sets for those systems.

[Scott Lurndal]

Al Kossow <a...@bitsavers.org> writes:
>On Thursday, June 1, 2023 at 4:25:06=E2=80=AFPM UTC-7, Al Kossow wrote:
>=20
>> I'm becoming concerned that what may have been produced about these syste=
>ms has been lost.
>
>I went back and looked at the implementations information I have for the 70=

>0/800/80/90 systems last night and

>they do run an interpreted system language on top of a microcoded 16 bit pr=

>ocessor. Even on the B1000 side with

>the internal memos that I have there was apparently no detailed documents o=

>n how the systems actually worked

>beyond the communicates and data structures which changed with each release=
>.
>I'm assuming if you had the need to know, you would just look at the source=

> code and none of this information was
>really ever released to customers.
>

>The situation is much worse for the non-B1000 small systems. I've not even =

>been able to find field service documentation
>beyond some schematic sets for those systems.

On the DCOM tape (http://bitsavers.org/bits/Burroughs/Vseries/Vseries_Tape_Collection.tar.gz)
there should be an NDL compiler that targets the B874 - takes a network topology
and station description and compiles it into the 'firmware' for the b874 (b800)
data communications processor which polls and selects stations and constructs
packets to send to the host (input from station) and packets sent from the host (station output).

[Timothy McCaffrey]

IIRC, the D-machine (in a non-B1000/B800 usage) was originally designed for the military
(ship board computer, I think). That may be why you are having a problem finding info
on it. You would think all of those systems would have been replaced decades ago...

- Tim