Graham Toal wrote:
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.
(apologies for cc: - I'm not sure whether you read list traffic
regularly or not!)
I've certainly got some old Acorn documentation with employees moaning
about the costs involved when it came to looking at licencing in the
early days of RISCiX; from memory (all that data's 4000 miles away right
now!) off-the-shelf alternatives to developing RISCiX were being
seriously considered.
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!)
I *might* have copies. I've certainly got some PANOS-related sources
back home (as you say, all in Modula-2). Won't be able to check until I
get back mid-November though. Those showed up on a disk just before I
left the UK, so I haven't had a chance to have a proper look through yet!
By the way I have copies of quite a few binaries for
this platform.
Oh boy. I'd *really* appreciate copies if you've got anything other than
the standard PANOS release binaries. I've been searching for anything
written for the ACW (or the 32016 beeb copro) for a few years now,
without much luck. QuickChip CAD was about the only thing I know of that
was ever run on the ACW which wasn't made by Acorn (other than student
assignments at Cambridge uni) - and it's not clear there whether it all
just ran on the B+ side of things anyway rather than being a 32016 port.
(I've got floppy install images, so can find out sometime when I have
the time and one of the ACWs out of storage!)
There was some 32016 code floating around at York uni years back
(1993-ish) and annoyingly I never backed up a copy :-( (I didn't have
the interest in preserving the less-common Acorn machines back then that
I do now)
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?
Well there are around ten ACWs left that I know of (or at least have
known of over the last few years; I'm not actively keeping track), which
isn't bad considering that only 50 or so seem to have been made way back
when! So PANOS is alive and well, just not very common :)
But yeah, from what I know of the instruction set's pretty easy for the
32016 (relevant manual should be on bitsavers, about 30MB IIRC). BeebEm
allows emulated coprocessors and of course is multi platform which is
always a good thing. How easy it is to add a new one I don't know
though, but presumably the relevant TUBE ULA emulation code should be in
the 6502 copro module which ships with BeebEm, so it's a case of
plugging 32016 code into that and you're away. I'll have a dig around
once I'm back in the UK though as someone was working on a 32016
emulator I believe a while ago - just can't remember who or the context,
but I'll have it in email archives back in the UK. Might be relevant.
cheers
Jules