NetNews Usenet Archive 1992 #27/NN_1992_27.iso/spool/comp/sys/misc/969

Internet Message Format | 1992-11-17 | 11.6 KB

Xref: sparky comp.sys.misc:969 alt.folklore.computers:16331
Path: sparky!uunet!usc!zaphod.mps.ohio-state.edu!magnus.acs.ohio-state.edu!usenet.ins.cwru.edu!agate!doc.ic.ac.uk!uknet!mucs!mshute
From: [e-mail] (Malcolm Shute)
Newsgroups: comp.sys.misc,alt.folklore.computers
Subject: PERQ Fanatics and Preservation Society
Message-ID: <[e-mail]>
Date: 16 Nov 92 17:20:08 GMT
Sender: [e-mail]
Followup-To: comp.sys.misc
Organization: Dept Computer Science, University of Manchester, U.K.
Lines: 224

Just over a month ago, I put out a request for information on users' groups,
and preservation societies for the PERQ workstation.

Well... I received an enormous amount of useful information, mainly from three
PERQ fanatics in the States, on how to get my machines up and running properly.

I also heard from several other PERQ owners.

Thank you, everyone who responded.

Rather than letting the information which I received, slip by through the
ether, I decided to edit it into an 'Alternative Help Manual'  (it contains
lots of useful stuff on what to do in various emergencies and problems).

As my contribution to a 'Summary to the Net'... here is Chapter One from
this document which I have managed to compile to date.  I hope that it is
of some interest to someone out there :-)

Cheers.

Malcolm.        v_@_

------------------------------------------------------------------------------
CHAPTER ONE

There appear to have been six models of the PERQ: the 1, 1A, 2, T2, 3 and  IV.
The highest estimates suggest that 4500 machines (of all types) were built and
sold.  They were standard issue for Science and Engineering  Research  Council
(SERC)  contracts in the UK, back in the early 1980s.  These were largely sup-
plied by the British company, International Computers  Limited  (ICL),  though
the machines appear to have been actually assembled in the USA.

It seems that disappointingly few of these machines are still around and work-
ing.   Few have found their ways into people's homes, as and when each one has
come up for scrapping.  A group of PERQ owners does exist,  though,  who  call
themselves the "PERQ Fanatics".  They are certainly now thinking of regrouping
as the "PERQ Preservation Society".  The material in this  document  has  been
provided  largely  by  them.   Explicitly, it includes text from the following
people (in alphabetical order):  Robert  Davis,  Duncan  Gibson,  Chris  Lamb,
Matthew  Laker,  Malcolm  Shute,  Bill  von  Hagen.  Email addresses, and some
snail-mail addresses, of these people, including that of the  editor  of  this
document, are listed next, the intension being that this report should contin-
ue to grow dynamically as and when new material is  contributed.   The  editor
is,  therefore,  eager to hear from anyone who would like to suggest any addi-
tions, corrections, or modifications.



1.1  Main present-day PERQ experts

In the USA, the only known users are Bob Davis, Chris Lamb,  Bill  von  Hagen,
and a few machines at a small company in Texas that are used for designing and
printing forms, and even they are in the process of phasing them out.

Robert Davis <[e-mail]> and <mystica![e-mail]>,  HAC,
Johns Hopkins University, Baltimore ([phone]).

Bill von Hagen <[e-mail]>, in Pittsburgh, used to work at CMU.  He re-
cently  purchased  the  final remains of Accent Systems from their bankruptcy.
As far as running systems go, he has two PERQ-1s, a  PERQ-2,  a  PERQ-T1,  and
four  PERQ-T2s.   Various  of  these  machines run PNX-1.0, POS-G5, Accent-S5,
PNX-5.2, POS-G6, Accent-S7, and he has one machine that runs  MPOS,  a  multi-
tasking version of POS that was written around 1982, but was never really com-
mercialized.

Chris Lamb <[e-mail]> owns three  PERQs,  and  runs  the  PERQ-Fanatics
mailing list (see below, in the next section).

Nick Felisiak <[e-mail]> Spider Systems Limited, Edinburgh,  UK  (+44
31 555 5166).  He bought the last two PERQ-T2s from the Department of Computer
Science at Manchester University (Man.CS).  He worked on PERQ at ICL  and  was
involved in PNX kernel and microcode.



1.2  Other present-day owners and users

Richard Wendland <[e-mail]> has a couple of PERQs.  He runs RSRE's
"Flex"  operating  system,  which is interesting but not Unix.  Praxis Systems
PLC, 20 Manvers Street, BATH  BA1 1PX, England.  (Tel: [phone],  Fax:
[phone] Groups 2&3).

Steve Romig <[e-mail]> worked at CMU when 3RRC  first  started
shipping PERQs.

Pete Sealey got a machine from King's College.

Steve Parkinson <[e-mail]> [ext 284] Ex-ICL  PERQ-3  s/w  development,
owns a (not yet working) PERQ-3.

Tony   Duell   <[e-mail]>,   actually   at   Bristol    University
<[e-mail]>,  intercepted a SERC PERQ.  He is currently near acquir-
ing a set of ciruit diagrams for the PERQ-T2, along with software,  and  other
information.

Robert Marshall <[e-mail]> owns a  PERQ-1A  from  Man.CS  running
PNX.   Ken Mayes <[e-mail]> owns a PERQ-1A from Man.CS running PNX.
Malcolm Shute <[e-mail]> (the editor of  this  document)  owns  two
PERQ-1As  from Man.CS running POS.  Department of Computer Science, Manchester
University, Oxford Road, Manchester, M13 9PL.  These four  machines  had  been
used up until the last moment for doing PCB and VLSI layout design.

At the same address, Matthew Laker <[e-mail]> owns a  PERQ-T2  from
Man.CS running PNX.

Also, David Alan Gilbert <[e-mail]> has a  PERQ-1A  on  permanent
loan  from  the  Manchester  Computer  Centre (MCC).  It seemed to be the only
remaining PERQ they had down there, and was used by the Archeological unit  or
something  similar  at  one point, and had been sitting in a back corridor for
three years.

In Australia, David Anderson was, when Bill von Hagen last spoke to him,  try-
ing to get his machine back up after a crash.



1.3  Mail distribution list

Posts for the entire group should be sent to "[e-mail]".  Chris
maintains  an archive of all the postings (and hopefully will soon be announc-
ing anonymous ftp access for Perq stuff).

To add or delete or correct an entry, or for  other  administrative  requests,
send  mail  to  "[e-mail]".   The Perq Fanatics mailing
list presently has the following members:

                 [e-mail],
                 [e-mail],
                 [e-mail],
                 [e-mail],
                 [e-mail],
                 [e-mail],
                 [e-mail]
                 [e-mail]


Advice on how to transport and install a PERQ that you have  just  managed  to
acquire, help on troubleshooting, documentation, spare parts, software and so-
cial events are all available in one form or another.  In the first  instance,
try contacting the manager of the mailing list, or the editor of this report.



1.4  History of the hardware, software, manufactures and users

The PERQ was specifically designed to do "Pascal Evaluation  Really  Quickly".
It  was originally designed and manufactured by the Three Rivers Computer Cor-
poration (3RCC), located in Pittsburgh, Pennsylvania, at the junction  of  the
Allegheny,  Mon  and Ohio Rivers.  3RCC later became the PERQ Systems Corpora-
tion, and later the Accent Systems Corporation.

They were truly the first commercial workstations (regardless of  some  design
and conceptual problems that accompany being the first-to-market).

When most people, especially in the UK, use the terms PERQ-1 and PERQ-2,  they
generally mean the PERQ-1A and PERQ-T2, respectively, since the former pair of
terms actually refer to two experimental machine types.


   PERQ-1          Original machine: 4K CPU, up to 1 Mbyte RAM, portrait
                   (sometimes referred to the PERQ-0, when 256K RAM
                   and a 12 Mbyte disk was standard)
   PERQ-1A         Original model: 16K CPU, up to 2 Mbyte RAM, portrait
                   or landscape.  Still used "IOB" for I/O and the
                   "OIO" board for Ethernet.
   PERQ-2          An horrendous rush-job to get the "new" machine
                   out the door.  8" disk interface and new cabinet
                   design, with up to 2 Mbyte memory, 16K CPU and either
                   monitor type.  Few of these were built, and the
                   8" drive controllers are infamous for their flakey
                   performance.
   PERQ-T2         The "real" PERQ-2, with the 16K CPU, up to 2 Mbyte
                   RAM, the EIO board with Ethernet built-in, 5.25"
                   disk support.  Probably, more of these were built
                   than any other model.
   PERQ-3          A completely different architecture.  Not the
                   "real" PERQ by any means.  It was going
                   to be a full 32-bit machine, and was actively in
                   development when the company went down.
   PERQ-IV         A "T2" with the 24-bit board set and backplane
                   modifications.  This allowed up to 4Mbyte of memory
                   (though with 24 address lines 16Mbyte is the theoretical
                   maximum, 4Mbyte memory boards seem to have been the
                   highest capacity offered.


Operating systems for the machines include: POS (PERQ Operating  System),  PNX
(a  sort  of  PERQ  Unix),  Accent.  The ACCENT operating system was initially
developed at Carnegie-Mellon University (CMU).

Regardless of its relatively youthful stature,  the  PERQ  has  a  significant
claim  on a place in history.  Moreover, in this business, 12 years (more like
13-14 now, if you include the fact that the design phase was underway in 1978)
is a very long time.

The PERQ was the first commercially available machine in  its  class.   Though
Xerox  built  the  Alto and other experimental machines, they were not offered
for sale to the public.  Brian Rosen, principal architect of the PERQ,  did  a
stint at Xerox PARC before returning to Three Rivers Computer, and if you com-
pare the architecture of the Xerox Alto and  the  PERQ,  you'll  see  striking
similarities.

The PERQ could be called the first "3M" machine: 1 Mips computing  power,  al-
most  1  Megapixel  display, and 1 Mbyte of RAM in a single-user system.  That
was a very important milestone in computing, and the PERQ  really  did  define
the  whole  industry.   Apollo  and  Sun and all the others that followed took
years to build a system around commercial micro-processor that  could  compete
with the PERQ for graphics performance.

The PERQ really did "push the envelope": for US$30,000 in  1980  you  had  the
most  powerful single-user computer available, with excellent graphics perfor-
mance on a 100 dpi display, a local 24 Mbyte disk, one Mbyte of memory,  sound
output,  RS232, GPIB, and Ethernet ports, and that in and of itself was an as-
tonishing feat.  Outside of the USA, ICL was pushing PNX, which  just  happens
to  be  the  first Unix with an in-kernel window manager, and distributed file
service, etc.  Inside the USA,  Accent  was  paving  the  way  in  distributed
operating  system (OS) research, with an elegant microkernel (they squeezed it
into 16K of writable control store (WCS), with two language  instruction  sets
and  room  to  spare)  and a new and powerful window manager (when X was still
Rev-10 and still a bogus hack floating around MIT).  PERQ "Spice Lisp" was the
first workstation implementation of Common Lisp.
--

Malcolm SHUTE.         (The AM Mollusc:   v_@_ )        Disclaimer: all