[9fans] Don't know much about history

[9fans] Don't know much about history

[Jack Johnson]

I like to comb second-hand stores for overlooked geek books, and came
across Interactive Programming Environments, which includes a copy of
Kernighan and Mashey's UNIX Programming Environment paper from '81. 
There's an interesting passage:

"UNIX has spawned a host of offshoots---at least six companies offer
or plan to offer systems derived from or compatible with the UNIX
system...."

With a footnote listing the companies:

"Cromemco, Onyx, Yourdon, Whitesmiths, Amdahl, and Wollongong Group."

Does anyone know what became of these companies and the products they
did (or didn't) produce?

-Jack

Re: [9fans] Don't know much about history

[Skip Tavakkolian]

> "Cromemco, Onyx, Yourdon, Whitesmiths, Amdahl, and Wollongong Group."
Cromemco: They built S-100 bus computers starting with z80 machines
and later (I think) using 68000 and 8086 (on VME machines?). I think
Cromix ran on the z-80 stuff and I thought they used UniSoft's port of
System V on 68000 versions.

Onyx: Z-8000 based computers.  Seem to recall they also had a V7
UNIX port of some sort running on them.

Yourdon: Told people how they should design software.

Whitesmiths: was started by P. J. Plauger (previously of Bell Labs),
had a UNIX workalike (V6, and later V7) called Idris, which didn't
require hardware MMU.  They had a line of nice C compilers for
a number of processors (starting with z-80 i think).  Got sold to
Intermetrics. 

Amdahl: made IBM 360/370 plug compatible (clone);
Was bought by Fujitsu (I think?)  I can't recall what UNIX variant
they had for it, or who did the work. This wasn't the MERT stuff, was it?

Wollongong Group:  They were the guys that had an IP stack
implementation for System V on the AT&T 3B line of
computers and the UNIX-PC.

Re: [9fans] Don't know much about history

[Richard Miller]

> Wollongong Group:  They were the guys that had an IP stack
> implementation for System V on the AT&T 3B line of
> computers and the UNIX-PC.

The Wollongong Group was a California company set up initially to
market the Interdata port of 7th edition Unix which had been done at
the University of Wollongong, Australia - hence the name.  The product
("Edition VII") was eventually adopted by Interdata (by then renamed
to Perkin-Elmer and again to Coherent) as a standard O/S offering on
their 32-bit minicomputers.  The Wollongong Group diversified into
other products (e.g.  Eunice = Unix emulator on VAX/VMS) and was
eventually bought by Attachmate.

Re: [9fans] Don't know much about history

[boyd, rounin]

> The Wollongong Group was a California company set up initially to
> market the Interdata port of 7th edition Unix which had been done at
> the University of Wollongong, Australia - hence the name. 

that sounds right.  strange that i've heard of them and 'the gong' being
close to sydney but then they evaporated.

wasn't the Interdata a 'strange' m/c?

maltby would probably know.

Re: [9fans] Don't know much about history

[Geoff Collyer]

The Interdata was odd.  I first tried to implement my interrupt
queuing scheme for changing fixed interrupt priorities (see the
Computing Systems paper) on one running Edition VII.  The machine
maintained a queue of pending interrupts in main memory.  Even with
the processor manual and the source code to Edition VII, I couldn't
get it to do what I wanted correctly.  I suspect that there was
something very subtle and not quite documented about the machine.  (I
did get my interrupt queuing scheme to work on the AMD 29200, another
odd machine.)  Steve Johnson has commented (see A Quarter Century of
UNIX) that he and dmr tried to get Interdata to fix various bugs that
they tripped across in their Unix port, without success.

Re: [9fans] Don't know much about history

[Richard Miller]

> Amdahl: made IBM 360/370 plug compatible (clone);
> Was bought by Fujitsu (I think?)  I can't recall what UNIX variant
> they had for it, or who did the work. This wasn't the MERT stuff, was it?

I've got a glossy A4 sheet here dated September 18 1980 which says
"Amdahl announces Universal Timesharing System Release 1.0 ... (UTS)"

Re: [9fans] Don't know much about history

[boyd, rounin]

> > Amdahl: made IBM 360/370 plug compatible (clone);
> > Was bought by Fujitsu (I think?)  I can't recall what UNIX variant
> > they had for it, or who did the work. This wasn't the MERT stuff, was it?

weren't fujitsu sued because that had stolen trade secrets from IBM?

Re: [9fans] Don't know much about history

[Richard Miller]

And then there was that other obscure software company ... quoting
from a contemporary brochure:

"Microsoft announces the XENIX Operating System, a 16-bit microprocessor
adaptation of Bell Laboratories' Version 7 UNIX system ... We're putting
the XENIX OS on the DEC PDP-11 [sic], Intel 8086, Zilog Z8000 and
Motorola 68000 ... The XENIX system's inherent flexibility, along
with this commitment from Microsoft, will make the XENIX OS the
standard operating system for the microcomputers of the eighties."

What a different world it could have been.

Re: [9fans] Don't know much about history

[Geoff Collyer]

Amdahl's machines were called 470s and their front-end communications
processors were called 4705s, if memory serves.  Their initial Unix
was called something like UTS/370 (I'm sure about the `UTS' part).  We
ran several virtual machines of UTS on an IBM 3033 that IBM donated,
for undergraduate use at U of Toronto in the early 1980s.  UTS was a
port of V7, with a few gratuitous C library interface changes, a port
of dmr's V6 PDP-11 C compiler, and no fsck.  Their fsck replacement
had the charming property that it only worked on file systems with a
handful of errors needing repair, so just when you really needed it,
it gave up.  I ported V7's fsck quickly and we used that instead.  The
system used ASCII internally and translated to and from EBCDIC where
necessary at the edges (some terminal I/O, printing, etc.).  It had
support for VM/370's virtual card readers, card punches and line
printers, so inter-virtual-machine communication was done via virtual
card images.  At the time, it seemed really fast.  It could fsck an
entire 3330 disk drive in 30 seconds.  `make' and `make -n' ran at
about the same speed.  Its Achille's heel was that IBM machines tended
to have operators, who thought they could just `force down' (crash)
virtual machines running UTS, though we told them over and over that
they needed to shut them down gracefully, including syncing them, and
provided them with a simple way to do so.

I'm pretty sure that UTS (d)evolved and become System V.something.

Google is unreachable now (as is much of the Internet), but I think
the Wollongong Group were best known for Eunice, a Unix emulator that
ran (slowly) on VAX/VMS.

Re: [9fans] Don't know much about history

[George Michaelson]

On Wed, 30 Jun 2004 00:58:54 -0700 Geoff Collyer <geoff@collyer.net> wrote:

 
>Google is unreachable now (as is much of the Internet), but I think
>the Wollongong Group were best known for Eunice, a Unix emulator that
>ran (slowly) on VAX/VMS.

in the absence of a real OS fork() with open FD 0 1 2 state, eunice beat
the hell out of the VMS alternatives for many things. I never liked
'mailboxes' which was their IPC model.

running eunice was a hoot!

-George

Re: [9fans] Don't know much about history

[boyd, rounin]

> in the absence of a real OS fork() with open FD 0 1 2 state, eunice beat
> the hell out of the VMS alternatives for many things. I never liked
> 'mailboxes' which was their IPC model.

story goes that sydney uni comp sci [basser] has the choice between
VMS and unix.  all those in the know wanted unix (some of 'em
ex-UNSW) and so a 60 user benchmark was run on an 11/780.

iirc, VMS swapped all over itself, smashing the 'system' pack.

unix didn't.

so basser got to run unix.

Re: [9fans] Don't know much about history

[boyd, rounin]

> ... and no fsck.

one time an ex-colleague screwed up and used the
wrong [new] kernel with an old [not upt to date]
RM03 [80mB; our root f/s].

the standalone tools worked-ish, but it took tape,
a screwdriver, a 9 track, adb and standalone fsdb.

we wasted a lotta 132x60 fanfold paper that night.

Re: [9fans] Don't know much about history

[boyd, rounin]

> Whitesmiths: was started by P. J. Plauger (previously of Bell Labs),
> had a UNIX workalike (V6, and later V7) called Idris, which didn't
> require hardware MMU.  They had a line of nice C compilers for
> a number of processors (starting with z-80 i think).  Got sold to
> Intermetrics. 

yes, Whitesmiths had a good reputation.

Re: [9fans] Don't know much about history

[Steve Simon]

The Perkin Elmer (né Interdata) 3210 was my introduction to
unix, (edition7) at college.

We had the source, all the source. We had the kernel, the
drivers, assembly of the diagnostic tapes, and the schematics.

A couple of the lecturers did a memory upgrade 1Mb to 4Mb,
with a soldering iron, a tube of rams, and  and a couple of
extra 74138 address decoders upside down on top of the original
ones.

happy days.

-Steve