Unix on BBC micro with 16032 coprocessor

Graham Toal gtoal at gtoal.com
Sun Oct 9 14:50:23 CDT 2005


> On Sat, October 8, 2005 6:32 pm, Pete Hollobon said:
>> Does anyone know if Xenix / Unix were ever available on the NS 16032
>> second
>> processor for the BBC micro? I remember reading that it was intended to be
>> made available in the user's guide years ago.

> It was supposed to be an option for the Cambridge Workstation since that
> also had a 16032/32016 2nd processor but I don't know if it ever made it
> into the wild. Graham Toal's the man who'll know I suspect, also Paul
> Williams (if he's around) and Jules R.

To the best of my recollection this was only ever talked about but
never implemented by anyone at Acorn.  If it was done externally I
have no knowlege of it.  I was one of the 3 or 4 people who argued
in favour of us selling the systems with Unix but Acorn was very much
a 'not invented here' company.  I think way back then there were
also significant costs associatated with licensing Unix, plus there
were few ports to archiectures other than Vax and 68000.

While we're on the subject of the NS processor... A couple of years ago
I got in touch with Peter Robertson, the author of Acorn's compilers for
that system (Pascal, C and Imp77) in the hope of getting the source of
the compiler from him.  Unfortunately he had not kept a copy of that code
generator.  I vaguely remember that there were sources escrowed somewhere
at the time.  Obviously the escrow will be long gone, but on the vague
offchance that someone from Acorn retained a copy... does anyone have the
source code for that compiler?  (Really just the back-end Icode to
binary Pass3, but _anything_ anyone has is wanted for the Edinburgh
Computer History Project!)

By the way I have copies of quite a few binaries for this platform.
Has anyone ever written an emulator for this architecture?  I haven't
found one.  It ought to be an easy one to write, it was a very regular
instruction set.  My guess for why there isn't one is that it was
never a very popular chip, but that hasn't stopped emulators being
written for lots of other obscure architectures!  I can imagine that
it wouldn't take much more than just a basic instruction set emulator
to make Panos live again, as the I/O could be done by emulating an
I/O processor Beeb and the tube chip.  There must be several Beeb
emulators around that can handle second processors to which we
could graft a NS emulator?

I think somewhere I have a paper listing of Mark Taunton's linker
for Panos.  It was a very well written piece of code.  I think it
was Acorn's only piece of Panos code (except for the compilers
themselves) that was written in Imp77.  Everything else was in
Modula II.  I also recently discovered Keith Rautenbach's & my
editor for Panos, the one which was a sort of EMACS-alike written
in ModII with a built-in mock LISP.  I have all the sources *except*
the LISP init files needed to start it up :-(  Also, the editor shot
some code over to the IO processor when it was invoked, which did some
extended keystroke handling - that part I do have...


A small aside: when I built my 6809 second processor for the Beeb,
while I was waiting for the chips to arrive I wrote a 6809 emulator
that ran on the 16032, written in Imp77.  It was faster than the
real chip!  Also it was my first emulator.  I wrote it on a 32016 2nd
proc that was on a Beeb with one of the early experimental Winchester
disks.  Of course the disk died, - about 2 days after I finished
the emulator - and the backup floppy was corrupt and the only source
I had was on paper.  I was too depressed to key it all in again
(and the paper listing was about a week from the final version)
and what's worse I lost the paper listing along with *all* my
historical computing papers when I emigrated from the UK to the
US 10 years ago :-/  Anyway, moral of this story was that it was
a very early wake-up call for me on the value of backups :-)  Ever
since then I have always had two hard drives containing everything
I've ever written, with one of them offline so they don't both
get taken out by the same lightning strike!  I am so glad that
disk drives are getting bigger faster than I can fill them - I've
always been able to keep everything I've ever done and just move
it over to the latest biggest drive when it becomes cheap enough
to afford.  I saw 500Gb hitachi drives at $260 today so I'm thinking
it's time to migrate again :-)


Graham
PS Joe Rigdon, the US Beeb you loaned me will be sent out in tomorrow's
mail.  Thanks very much for letting me use it!



More information about the cctech mailing list