The ELLIOTT 503
e_erp...@hotmail.com
Elliott 803 computer that I found in the book
CURRENT COMPUTERS 1969.
In this posting, I will provide the data that this
book has about the Elliott 503, a more advanced
version of the 803.
Does anyone have any stories about programming or
using either the 503 or 803?
It would be also be interesting to know if anyone
still has the programming manuals for these computers.
Considering how most computer documentation gets
trashed, it would actually be amazing if more than
a few manuals are still about.
From CURRENT COMPUTERS 1969.
COMPUTING SYSTEM: ELLIOTT 503
General Characteristics:
========================
The 503 is the most powerful in the present range of the Elliott
solid state computers. While primarily designed for scientific
work it is fully adaptable to commercial applications. It is very
fast and is capable of controlling a large range of peripheral devices.
First Installed: 1963
Present Availability: 6 months (1966)
Country of Manufacture: Great Britain
Selling Price: £56,000 to £416,000
Rental per month:
£1,017 to £9,200
Amount of initial charge: Nil
Purchase Option conditions: None
Installation cost: Included in price.
Carriage: Included in price
Warranty: 6 months
Physical Characteristics:
=========================
Floor area average complete system: 600 sq. ft.
Maximum floor loading: 250 lbs. per sq. ft.
Power requirements: 9.2 - 35 kva.
Air conditioning requirements:
70 +/- 4 deg. F, 40-60% RH, 95% efficient @ 10 microns
Central Processing Unit Used: 503
Special features which can be added
to Central Processing Unit: Interface Matching Unit
Operation Times:
================
Addition (fixed point): From 7.2 usecs.
Multiplication (fixed point): From 34.8 to 46.8 usecs.
Division (fixed point): From 68.7 usecs.
Addition (floating point): From 13.2 to 30.3 usecs.
Multiplication (floating point): From 31.5 to 41.0 usecs.
Division (floating point): From 60.9 to 61.8 usecs.
Cycle Time:
Logic ) 3.6 usecs
Storage) 3.6 usecs
Word Length: 39 bits including sign
Basic Storage:
Type: Ferrite Core
Size: 8192 words
Additional Storage:
Type: Core
Size: 131K
Input/Output:
=============
Type: Magnetic Tape Handlers
Speed: 60,000 ch/sec.
Type: Line Printers
Speed: 330/1250 l.p.m.
Type: Magnetic Film Handlers
Speed: 4300 c.p.s.
Type: Card Read/Punch
Speed: 300 c.p.m.
Type: Card Reader
Speed: 340 c.p.m.
Type: Graph Plotter
Speed: 3" per sec, accuracy .01"
Type: Typewriter
Speed: 10 c.p.s.
Type: Paper Tape Reader
Speed: 1000 c.p.s.
Type: Paper Tape Punch
Speed: 100 c.p.s.
Software Available:
===================
Elliott Autocode
Algol
Symbolic Assembly Code
Fortran to Algol Translator
Input/Output System
Simulator Language
Time Sharing
Fortran IV
over 200 library and complete application programmes.
System Analyst Service: Not available
Programming Service: Not available
General Comments:
=================
Same instruction code as 803 but speed of operation
is nearly 100 times greater. There are a number of interrupt
lines which not only permit running of programmes in
parallel but also are used in connection with on-line operation
of the 503.
Selling Organisation in Great Britian:
======================================
Elliott Automation Computers Ltd.
Sent via Deja.com http://www.deja.com/
Before you buy.
Eric v/d Meer
wrote:
> In a previous posting, I gave the detail about the
> Elliott 803 computer that I found in the book
> CURRENT COMPUTERS 1969.
>
> In this posting, I will provide the data that this
> book has about the Elliott 503, a more advanced
> version of the 803.
>
> Does anyone have any stories about programming or
> using either the 503 or 803?
>
> It would be also be interesting to know if anyone
> still has the programming manuals for these computers.
>
> Considering how most computer documentation gets
> trashed, it would actually be amazing if more than
> a few manuals are still about.
>
I still have fairly complete set of manuals of the 503, including SAP,
the library routines, and even Algol3. If you're interested, I could
make a list.
Eric
e_erp...@hotmail.com
Eric v/d Meer <eric.va...@net.hcc.nl> wrote:
[snip]
> > Does anyone have any stories about programming or
> > using either the 503 or 803?
> >
> > It would be also be interesting to know if anyone
> > still has the programming manuals for these computers.
> >
> > Considering how most computer documentation gets
> > trashed, it would actually be amazing if more than
> > a few manuals are still about.
> >
>
> I still have fairly complete set of manuals of the 503, including SAP,
> the library routines, and even Algol3. If you're interested, I could
> make a list.
>
> Eric
>
Eric,
That would be great! I would be interested in seeing such a list.
--Eric
Werner Uelpenich
artikle which mentioned that Arch 1000 is equivalent to Elliot 503.
The Arch 1000 was used to control a chemical plant at the end of the
60ties. It had retired when I worked at it as student 1971. It had 3 lagre
cabinets, the middle one was the central processing unit, the left one
housed some I/O interfaces (papertape reader, Siemens Fernschreiber (German
teletype) interface, stepping motor outputs), and the right one contained
an analog to digital converter, an amplifier for thermo couple instruments
and 200 mercury wetted relais as multiplexor.
It was a 18 bit cpu, 2 blocks of 4k Words core memory, a small diode matrix
of ~96 words hard wired monitor (operating system). It had an instruction
set of 13 instructions (I think these 13 instructions are the absolute
minimum to do all tasks, a real RISC), Switch register, speaker. the clock
speed was 100 kHz.
The diagrams were dated from 1963, I think that machine was build 1966. The
documentation had a complete set of diagrams.
The logic was build of "minilogs". A minilog was a small plastic case with
2 germanium transistors and some resistors. Most of them contained two NOR
gates, each of them with 3 inputs via resistor and 1 expansion input
directly connected to the basis of the transistor. very often these two NOR
gates were combined to a flop-flop.
Who remember this Arch 1000 and is it really equivalent to Elliot 503?
Werner
e_erp...@hotmail.com schrieb:
> In a previous posting, I gave the detail about the
> Elliott 803 computer that I found in the book
> CURRENT COMPUTERS 1969.
>
> In this posting, I will provide the data that this
> book has about the Elliott 503, a more advanced
> version of the 803.
>
> Does anyone have any stories about programming or
> using either the 503 or 803?
>
> It would be also be interesting to know if anyone
> still has the programming manuals for these computers.
>
> Considering how most computer documentation gets
> trashed, it would actually be amazing if more than
> a few manuals are still about.
>
Eric v/d Meer
wrote:
> In article <240520002340091406%eric.va...@net.hcc.nl>,
> Eric v/d Meer <eric.va...@net.hcc.nl> wrote:
> [snip]
>
> > > Does anyone have any stories about programming or
> > > using either the 503 or 803?
> > >
> > > It would be also be interesting to know if anyone
> > > still has the programming manuals for these computers.
> > >
> > > Considering how most computer documentation gets
> > > trashed, it would actually be amazing if more than
> > > a few manuals are still about.
> > >
> >
> > I still have fairly complete set of manuals of the 503, including SAP,
> > the library routines, and even Algol3. If you're interested, I could
> > make a list.
> >
> > Eric
> >
>
> Eric,
>
> That would be great! I would be interested in seeing such a list.
>
> --Eric
>
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.
Eric,
OK, here goes.
Following are the highlights of the 503 TECHNICAL MANUAL set. It came
in four volumes, where volume 2 was split in two sub-volumes due to
size limitations of the binders.
VOLUME 1 503 FUNCTIONAL SPECIFICATION
======================================
CONTENTS LIST
PART 1 GENERAL DESCRIPTION
Section 1 - Description of components of the 503 system
Section 2 - General data
Section 3 - Installation details
PART 2 BASIC 503 SYSTEM
Section 1 - Central processor
The Initial Instructions
The Reserved Area
The Registers
The Instruction Code
B-modification
8-Channel Paper Tape Code
5-Channel Paper Tape Code
Section 2 - Control Station
Paper Tape Readers
Reader Control Panel
Paper Tape Punches
Punch Control Panel
Control Typewriter
Buttons and Lamps on the Main Control Panel
Section 3 - Interrupt facilities
Section 4 - Operation of interrupts
Section 5 - Error detection
PART 3 PERIPHERAL CONTROL
Section 1 - Autonomous transfer
The 76 Instruction
The 77 Instruction
Tagging
PART 4 PERIPHERAL DEVICES
Section 1 - Backing store
Section 2 - Magnetic tape
Section 3 - Line printer
Section 4 - Card reader/punch
Section 5 - Direct input and output
Section 6 - Elliott card reader
Section 7 - Digital incremental plotter
Section 8 - Interface Matching Unit
Section 9 - High Speed Character Printer
Section 10 - Magnetic Film
Section 11 - On-Line Clock
PART 5 OFF-LINE EQUIPMENT
Section 1 - Flexowriter
VOLUME 2: PROGRAMMING INFORMATION
==================================
CONTENTS LIST
INTRODUCTION
PART 1: PROGRAM SYSTEMS
Section 1: Basic Machine Programming
Section 2: Symbolic Assembly Mark 1
Section 3: 503 Algol Mark 1
Section 4: 503 Autocode
Section 5: ALGOL Mk. 3
PART 2: 503 LIBRARY PROGRAMS
Section 1: Reserved Area Program
Section 2: General Information on the Library
Section 3: Program Specifications (excluding test programs)
[The library consisted of some 50 programs and subroutines
with various purposes. The two main groups were mathematical
functions and utilities and subroutines for accessing
peripheral devices.]
Section 4: Daily Test Program Specifications
PART 3: THE INTEGRATED SOFTWARE SYSTEM FOR THE NON-BASIC 503
Section 1: Introduction
PART 4: PROGRAM LANGUAGES FOR THE NON-BASIC 503
Section 1: Symbolic Assembly Mark 2
Section 2: 503 ALGOL Mark 2
Section 3: FORTRAN IV
PART 5: PROGRAMMING AND OPERATING AIDS FOR THE NON-BASIC 503
Section 1: Operational Techniques
Section 2: The Storage Planning and Allocation System
Section 3: Peripheral Control Program
Section 4: The Controlling Programs
Section 5: Segmented Tape Administrative Routines
[This is the advanced batch operating system that
used magnetic tape for input and output spooling.
It was never delivered.]
VOLUME 3: OPERATING INFORMATION
================================
VOLUME 4: ENGINEERING MAINTENANCE
==================================
As you may gather from the lack of a contents list, volumes 3 and 4 are
missing.
Howevever, there is also Elliott Facts, a tiny booklet with the
information that a programmer would need every day, such as the
commands for the Reserved Area Program, the instruction set and the
powers of 2 up to 39.
Cheers
Eric
Harold Rabbie
<e_erp...@hotmail.com> wrote in message
news:8gcsq3$mgi$1...@nnrp1.deja.com...
> In a previous posting, I gave the detail about the
> Elliott 803 computer that I found in the book
> CURRENT COMPUTERS 1969.
>
> using either the 503 or 803?
On a grammar school field trip to the Elliott Automation factory in Welwyn
Garden City, they taught us some basic Algol-60. I wrote my first programme
on a 5-track paper tape using a Flexowriter. The operator fed the tape into
a reader, and then used the console switches to load the various phases of
the compiler, linker, and loader from a magtape. The standard output was a
paper tape punch, so you took your output tape and printed it with the
Flexowriter. ISTR they showed us some delay-line memory devices, which
contained a long coil of piano wire with piezo transducers at each end that
acted like a sonic shift-register. Jeez, that must have been in the early
'60s, and I'm still writing programs nearly 40 years later. But I don't use
Baudot code or Algol any more....
Harold Rabbie
Los Gatos, California
Phil Howell
at Portsmouth Polytechnic in about 1971/72 there was an ICT 4100
which (i think) was an Elliot in disguise. The languages were algol and
fortran. The algol compiler also had the directive $NEAT so you could embed
assembler code for low-level tinkering.
The acronym neat stood for National Elliot Assembly Technique which was some
sort of collaboration with NCR who also marketed the machines, but I don't
know what they called them.
Phil
<e_erp...@hotmail.com> wrote in message
news:8gcsq3$mgi$1...@nnrp1.deja.com...
> In a previous posting, I gave the detail about the
> Elliott 803 computer that I found in the book
> CURRENT COMPUTERS 1969.
>
> In this posting, I will provide the data that this
> book has about the Elliott 503, a more advanced
> version of the 803.
>
> Does anyone have any stories about programming or
> using either the 503 or 803?
>
> It would be also be interesting to know if anyone
> still has the programming manuals for these computers.
>
> Considering how most computer documentation gets
> trashed, it would actually be amazing if more than
> a few manuals are still about.
>
Charles Richmond
>
> Some Elliot machines were re-badged.
> at Portsmouth Polytechnic in about 1971/72 there was an ICT 4100
> which (i think) was an Elliot in disguise. The languages were algol and
> fortran. The algol compiler also had the directive $NEAT so you could embed
> assembler code for low-level tinkering.
> The acronym neat stood for National Elliot Assembly Technique which was some
> sort of collaboration with NCR who also marketed the machines, but I don't
> know what they called them.
>
But NCR had a business language called NEAT. I was told it was similar
to COBOL. Has anyone here used the NCR NEAT language??? What was it
like??? Please, someone post some details...
--
+-------------------------------------------------------------+
| Charles and Francis Richmond <rich...@plano.net> |
+-------------------------------------------------------------+
Brian Boutel
Charles Richmond wrote:
>
> Phil Howell wrote:
> >
> > Some Elliot machines were re-badged.
> > at Portsmouth Polytechnic in about 1971/72 there was an ICT 4100
> > which (i think) was an Elliot in disguise. The languages were algol and
> > fortran. The algol compiler also had the directive $NEAT so you could embed
> > assembler code for low-level tinkering.
There was a NCR Elliot 4100. ICT would have re-badged a lawnmower if
they could make a buck out of it.
> > The acronym neat stood for National Elliot Assembly Technique which was some
> > sort of collaboration with NCR who also marketed the machines, but I don't
> > know what they called them.
> >
> I do *not* know about the Elliot machines, or how they were sold by NCR.
> But NCR had a business language called NEAT. I was told it was similar
> to COBOL. Has anyone here used the NCR NEAT language??? What was it
> like??? Please, someone post some details...
>
I worked for NCR in UK in 1962-63, working initially with the Elliot
803. The NCR 315 arrived around the end of '62. NEAT was the assembly
language for the 315, which was not an Elliot design, although Elliots
manufactured them in UK for NCR. "NEAT" also became a marketing term
associated with other system products.
The Cobol-like language was probably BEST, but could have been Language
H.
The 315 had a 12-bit word (known as a "slab"). This could be seen
(depending on the instruction used) as either 2 x 6-bit characters or 3
x 4-bit bcd digits. I seem to remember devising a way of extracting some
bits from slab by using a mixture of digit and character instructions.
There were no conventional bit operations, or even shifts.
The main claim to fame for the 315 was the CRAM storage. A removable
cartridge which (in the original version) held 127 cards on 7
semicircular rods. The cards had sock-shaped cutouts, arranged so that
turning the rods would cause one card to drop from the supporting rods
onto a rotating drum, where it could be read, after which the vacuum
holding it to the drum was released, and it returned up the chute to its
place. This process was accompanied by a very loud noise. Each card had
7 data tracks with about 1500 characters on each. Later versions
increased the capacity, bit the whole scheme was too complicated and
unreliable compared to the disk drives with removable packs which
inevitably replaced it.
--brian
Andrew Gabriel
Charles Richmond <rich...@dallas.net> writes:
>Phil Howell wrote:
>>
>> Some Elliot machines were re-badged.
>> at Portsmouth Polytechnic in about 1971/72 there was an ICT 4100
>> which (i think) was an Elliot in disguise.
>>
The 4100 was an NCR machine, which Elliott made under licence.
When Elliott (and the other related computer manufacturers
which were aquired by GEC) reorganised around 1970, the 4100
was classed as a Data Processing computer, and therefore went
to ICT (which later became ICL). The Real Time computers went
to GEC Computers.
See diagram at bottom of web page
http://www.cucumber.demon.co.uk/geccl/19471972/index.html
which are old presentation slides produced at GEC Computers'
formation, to explain to staff and customers how the product
range and companies were being split.
--
Andrew Gabriel
Consultant Software Engineer
Eric v/d Meer
<werner.u...@netcologne.de> wrote:
> I've got my first programming experiance with an Arch1000. I found an
> artikle which mentioned that Arch 1000 is equivalent to Elliot 503.
>
> The Arch 1000 was used to control a chemical plant at the end of the
> 60ties. It had retired when I worked at it as student 1971. It had 3 lagre
> cabinets, the middle one was the central processing unit, the left one
> housed some I/O interfaces (papertape reader, Siemens Fernschreiber (German
> teletype) interface, stepping motor outputs), and the right one contained
> an analog to digital converter, an amplifier for thermo couple instruments
> and 200 mercury wetted relais as multiplexor.
>
> It was a 18 bit cpu, 2 blocks of 4k Words core memory, a small diode matrix
> of ~96 words hard wired monitor (operating system). It had an instruction
> set of 13 instructions (I think these 13 instructions are the absolute
> minimum to do all tasks, a real RISC), Switch register, speaker. the clock
> speed was 100 kHz.
>
> The diagrams were dated from 1963, I think that machine was build 1966. The
> documentation had a complete set of diagrams.
> The logic was build of "minilogs". A minilog was a small plastic case with
> 2 germanium transistors and some resistors. Most of them contained two NOR
> gates, each of them with 3 inputs via resistor and 1 expansion input
> directly connected to the basis of the transistor. very often these two NOR
> gates were combined to a flop-flop.
>
> Who remember this Arch 1000 and is it really equivalent to Elliot 503?
>
> Werner
Werner,
I don't think the Arch 1000 was much like the 503. With a word length
of 18 bits it was more like the 903. Its orderocde may owe something to
the 503's, however.
HTH
Eric v/d Meer
>
> e_erp...@hotmail.com schrieb:
e_erp...@hotmail.com
Eric v/d Meer <eric.va...@net.hcc.nl> wrote:
[snip]
> > >
> > > I still have fairly complete set of manuals of the 503, including
SAP,
> > > the library routines, and even Algol3. If you're interested, I
could
> > > make a list.
> > >
> > > Eric
> > >
[snipping a most interesting table of contents]
Thanks for providing that info Eric!
Someday I hope someone puts up the programming manual for the
Elliott 803/503 computers. I think there is enough interest
to justify it being on a webpage.
--EE
Rick Lugg
news:392E5EE7...@dallas.net...
> Phil Howell wrote:
> >
> > Some Elliot machines were re-badged.
> > at Portsmouth Polytechnic in about 1971/72 there was an ICT 4100
> > fortran. The algol compiler also had the directive $NEAT so you could
embed
> > assembler code for low-level tinkering.
some
> > sort of collaboration with NCR who also marketed the machines, but I
don't
> > know what they called them.
> >
> to COBOL. Has anyone here used the NCR NEAT language??? What was it
> like??? Please, someone post some details...
>
NCR badged and sold the Elliott 4100 (as the NCR-Elliott 4100) and NCR
provided the OS, some peripherals (tapes, CRAM, printers etc.) as options.
Languages were NEAT 4100 (National Electronic Autocode Technique) which was
really a macro-assembler. I think we also had a Cobol compiler and Algol.
From my memory, the 4120 processor was a 24bit word, 16K memory. The 4130
was a further development and would have had an NCR multi-processing OS. I
think cycle time for that was 1.2 micro-seconds and started at 32K memory up
to 128K !
Customers at the time included Municipal authorities, PIDA, Cater Ryder, and
many more.
The NCR machine at the time (1966-) was the NCR 315 which was just too
expensive for much UK use. Mercantile Credit had a system which had been
built (or more likely assembled) by Elliott in Borehamwood for NCR. There
were other users of 315 (I think about 6 in the UK, but I can't remember
them!). In the early 60s, Mercantile were running concurrent despooling
whilst other processes were running, and could provide a "settlement" figure
for a loan account through teletypes in regional offices doing an on-line
lookup to a summary file. Spool files, the summary (from yesterday) and the
masterfile update processes would all run concurrently using CRAM units.
CRAM stands for Card Random Access Memory (though I know some claimed it
stood for Chaos Reaches All-time Magnitude). It was one of the first true
random access devices with changeable media magnetic cards held on rotatible
spindles that allowed one specific card (supposedly) to drop onto a drum. It
then became a magnetic drum for a while till the card was returned to the
deck. The 315 used NEAT 315 as its assembler. When the NCR Century was
introduced, it had a much wider range of models and with the absorption of
Elliott into ICL made the 4100 range less attractive to market.
NCR also offered BEST (Business EDP System Technique) which was a forms
based questionnaire for developing programs for 315. Basically you filled
in forms about the input transaction data and the outputs needed and a
program was generated. That capability migrated into services called "Major
Functions" in NEAT/3.
The Century introduced NEAT/3 which was a three level compiler/assembler
which had CoBOL like high level commands and shorthand data definitions,
plus lower level macro assembler capabilities. The third level was the
"protected" level of services that were supposed to be available to OS
developers only. CoBOL was also available on Century and I recall seeing
RPG though I don't know anyone who used it on the machines.
--
Rick Lugg
Werner Uelpenich
http://www.cucumber.demon.co.uk/geccl/19471972/index.html
where Elliot 503 and Arch 1000 is in the same box 1963 which means that they were
first produced in the same year.
Werner
Eric v/d Meer
wrote:
Eric,
Maybe something could be arranged. I don't have a website, though.
Furthermore I suspect there may be copyright issues.
BTW I've tried to send you email but it bounces.
Eric
Andrew Gabriel
Werner Uelpenich <werner.u...@netcologne.de> writes:
>I found the source of the information. it was
>http://www.cucumber.demon.co.uk/geccl/19471972/index.html
>where Elliot 503 and Arch 1000 is in the same box 1963 which means that they were
>first produced in the same year.
I put that table together from lots of sources, including records from
GPT at Borehamwood (one of Elliott company's grandchildren) who still
had a few Elliott records. The dates might well not all be measuring
the same thing - i.e. press launch verses first shipped systems, as that
level of detail was no longer available. It would be difficult to claim
definately that these two machines were produced in the same year, but
it is quite likely that if this isn't the case, there's probably no more
than a year's difference in them.
I am in contact with a number of people who worked for various parts of
the company during the 1960's (mostly retired now of course), and I could
try asking next time I see them if anyone can recall any further details.
Dave Daniels
Rick Lugg <rick...@knickers.iname.com> wrote:
> really a macro-assembler. I think we also had a Cobol compiler and Algol.
> From my memory, the 4120 processor was a 24bit word, 16K memory. The 4130
There was a FORTRAN IV compiler as well. I cut my programming
teeth on it during the mid-seventies. One of the UK universities
(York, I think) did a Basic compiler too.
Dave Daniels
--
ANTISPAM: Please note that the email address above is false. My
correct address is:
dave_daniels<at>argonet<dot>co<dot>uk
Please replace the <at> and <dot>s with @ and . respectively when
replying - Thanks!
Brian {Hamilton Kelly}
> Does anyone have any stories about programming or
> using either the 503 or 803?
I could probably remember a few: I could also remember using the Elliott
402, which was a three-and-a-half-address machine, where one had to
optimize division instructions in the same manner as for the EE Deuce, so
that the drum would be in the right place to pick up the next instruction
on completion of the divide op.
> It would be also be interesting to know if anyone
> still has the programming manuals for these computers.
>
> Considering how most computer documentation gets
> trashed, it would actually be amazing if more than
> a few manuals are still about.
All my programming manuals, for about forty different machines[1],
including the 503/803, were donated to the Science Museum in London.
AIUI, they now form "The Hamilton Kelly Archive", and are available for
perusal by scholars.
[1] Earliest machine was the Ferranti Mk.1*, of 1948 vintage.
--
Brian {Hamilton Kelly} b...@dsl.co.uk
"We have gone from a world of concentrated knowledge and wisdom to one of
distributed ignorance. And we know and understand less while being incr-
easingly capable." Prof. Peter Cochrane, BT Labs
Brian {Hamilton Kelly}
In article <392D8248...@netcologne.de>
werner.u...@netcologne.de "Werner Uelpenich" writes:
> Who remember this Arch 1000 and is it really equivalent to Elliot 503?
The various Arch series (of which the 1000 was only one) were made by
Elliott Automation for process control (somewhat akin to the various
Ferranti Argus series). As you say, they were built around minilog units
(sort of primitive disrete-component "integrated circuits").
The CPU was a completely different architecture to the 803/503 range; as
you say it used 18-bit words (IIRC, there was another Arch series that
used 12-bit words), whereas the 803/503 were general purpose computers
with 39-bit words.
Elliotts were in the forefront of the whole concept of industrial
automation: they probably sold more machines for that area than they did
for GP work.
Brian {Hamilton Kelly}
how...@snowyhydro.com.au "Phil Howell" writes:
> Some Elliot machines were re-badged.
> at Portsmouth Polytechnic in about 1971/72 there was an ICT 4100
> which (i think) was an Elliot in disguise. The languages were algol and
> fortran. The algol compiler also had the directive $NEAT so you could embed
> assembler code for low-level tinkering.
The 4100 series of machines (of which there were two, the 4120 and the
4130) were made by Elliotts towards the late 60s. They were based upon a
24-bit word (which those of us that had been used to the 803/503 39-bit
word found a great encumbrance when porting over some ballistics
programs).
The 4100 series was retained by ICT after the formation of the latter
from ICT (which had the 1900 series of computers), from English Electric
(which had the System 4 series, which was tied in with an American
company, can't remember which), and Elliott-Automation (which had the
4100 series).
Of these three architectures, the most technically advanced was the
System 4, next the 4100, and finally the 1900. In an amazing parallel to
VCR formats, the first to be killed off was System 4 (akin to the
Phillips/Grundig V2000 format), then the 4100 (cf. Betamax), leaving the
technically inferior 1900 (cf. VHS) as the only remaining product.
This short-sighted-NIH-ness of ICT within ICL meant that it was too long
before they attempted to be innovatory with the 2900 series, and led to
the total demise of the British computer industry, IMNSHO.
> The acronym neat stood for National Elliot Assembly Technique which was some
> sort of collaboration with NCR who also marketed the machines, but I don't
> know what they called them.
Before the formation of ICT, the machines (as with all other products of
the NCR/Elliott symbiosis (see other post)) were sold as the "Elliott-NCR
4120" (or 4130). Incidentally, my collection of over twenty-five manuals
for the 4100 series is also in the Hamilton Kelly Archive at the Science
Museum. Inter alia, this includes the manual for NICE: Normal Input and
Control Executive (think JCL).
Brian {Hamilton Kelly}
a__fake_...@127.0.0.1 "Dave Daniels" writes:
> In article <8gtd7g$son$1...@ctb-nnrp2.saix.net>,
> Rick Lugg <rick...@knickers.iname.com> wrote:
> > really a macro-assembler. I think we also had a Cobol compiler and Algol.
> > From my memory, the 4120 processor was a 24bit word, 16K memory. The 4130
>
> There was a FORTRAN IV compiler as well. I cut my programming
> teeth on it during the mid-seventies.
FORTRAN on the earlier 803/503 was "interesting". IIRC, an Australian
university insisted upon a compiler for this language, which Elliott's
had never bothered with. They got CAP (who'd written the Algol 60
compiler) to provide one; it actually translated FORTRAN into Algol60,
before compiling the latter!
> One of the UK universities
> (York, I think) did a Basic compiler too.
That would probably have been Terry Froggett; I remember his working on
the Basic for the 903/905 series, and York definitely had a 4130.
Brian {Hamilton Kelly}
and...@cucumber.demon.co.uk "Andrew Gabriel" writes:
> In article <392E5EE7...@dallas.net>,
> Charles Richmond <rich...@dallas.net> writes:
> >Phil Howell wrote:
> >>
> >> Some Elliot machines were re-badged.
> >> at Portsmouth Polytechnic in about 1971/72 there was an ICT 4100
> >> which (i think) was an Elliot in disguise.
> >>
> >I do *not* know about the Elliot machines, or how they were sold by NCR.
>
> The 4100 was an NCR machine, which Elliott made under licence.
I'm pretty sure that it was only ever an Elliott machine, but sold
through NCR's superior marketing organization. They were certainly all
made at Borehamwood; I rmember wandering about the shop floor seeing
dozens of machines under construction --- never before had I ever seen so
many computers.
Charles Richmond
>
> [snip...] [snip...] [snip...]
>
> FORTRAN on the earlier 803/503 was "interesting". IIRC, an Australian
> university insisted upon a compiler for this language, which Elliott's
> had never bothered with. They got CAP (who'd written the Algol 60
> compiler) to provide one; it actually translated FORTRAN into Algol60,
> before compiling the latter!
>
"There was a hunter who had a beautiful, titanium tiger gun with
gold inlays and marvelous stock carvings. A group of reports
ask him about it and the hunter says:
"Well, if I see an African tiger, *bang!*...I shoot it with my tiger gun!
And if I see an Indian tiger, *bang!*...I shoot it with my tiger gun!"
A reporter asks: "But what if you meet up with an elephant???"
The hunter replies: "I paint stripes on it and shoot it with my tiger gun!"
</Joke>
So they painted stripes on the FORTRAN, and shot it with the Algol60 compiler.
IMHO this technique is often used in programming...something is an odd-man-out
type of construct, you doctor it up to look like something else, and run it
through the regular code you already have.
When Seymour Cray wrote the FORTRAN compiler for the CDC-1604, he wrote
it as
a series of assembly language macroes. This allowed it to compile quickly,
although the run-time ended up being slow because of this technique. So he
was sort of painting stripes on it and shooting it with the tiger gun...
(P.S. Yes...I know that tigers and elephants are endangered species. But
these are "virtual" tigers and elephants. You can pull those out of your
hat any old day...)
Charles Richmond
>
> [sinp...] [snip...] [snip...]
>
> Before the formation of ICT, the machines (as with all other products of
> the NCR/Elliott symbiosis (see other post)) were sold as the "Elliott-NCR
> 4120" (or 4130). Incidentally, my collection of over twenty-five manuals
> for the 4100 series is also in the Hamilton Kelly Archive at the Science
> Museum. Inter alia, this includes the manual for NICE: Normal Input and
> Control Executive (think JCL).
>
>
> "We have gone from a world of concentrated knowledge and wisdom to one of
> distributed ignorance. And we know and understand less while being incr-
> easingly capable." Prof. Peter Cochrane, BT Labs
>
of dumb machines to dumb people in front of smart machines."
Brian Boutel
>
>
>
> The 4100 series was retained by ICT after the formation of the latter
> from ICT (which had the 1900 series of computers), from English Electric
> (which had the System 4 series, which was tied in with an American
> company, can't remember which), and Elliott-Automation (which had the
> 4100 series).
>
The ICT 1900 was based on the Ferranti-Packard FP6000.
System 4 was based on the RCA Spectra 70, which looked like an IBM 360.
>
> This short-sighted-NIH-ness of ICT within ICL meant that it was too long
> before they attempted to be innovatory with the 2900 series, and led to
> the total demise of the British computer industry, IMNSHO.
>
It is strange to criticise ICT/ICL about short-sighted-NIH-ness. Can you
remember the last machine they actually designed themselves, as opposed
to buying either the design or the company, before the 2900?
Back in 1964 the company I was with (in UK) were looking to buy a
machine after using IBM service bureaux. The boss said that on no
account was I to approach him with a proposal to buy ICT. He had had too
many bad experiences with them. We decided on a Honeywell 200. At that
point IBM tried their usual nasty tricks. They approached the Chairman
warning him about the foolishness of such a course of action. He had the
good sense to tell them that he employed people to make that kind of
decision for him.
Later, (early '70s), ICL tried a similar trick here in New Zealand. The
Govt decided to fund mainframes for a the NZ universities, on a single
contract. The decision was to buy Burroughs 6700s, but ICL approached
the Govt and said that we would be wise to reconsider and buy British
(i.e ICL 2900s) if NZ wanted to get any concessions in the negotiations
with Britain over UK joining the European Common Market. Not exactly a
move calculated to impress, and the B6700s were bought. Later, the NZ
Govt did buy a 2980, though.
--brian
Peter Onion
b...@dsl.co.uk (Brian {Hamilton Kelly}) writes:
> [Once again, apologies for the late response.]
>
> In article <392D8248...@netcologne.de>
> werner.u...@netcologne.de "Werner Uelpenich" writes:
>
>> Who remember this Arch 1000 and is it really equivalent to Elliot 503?
>
> The various Arch series (of which the 1000 was only one) were made by
> Elliott Automation for process control (somewhat akin to the various
> Ferranti Argus series). As you say, they were built around minilog units
> (sort of primitive disrete-component "integrated circuits").
>
> The CPU was a completely different architecture to the 803/503 range; as
> you say it used 18-bit words (IIRC, there was another Arch series that
> used 12-bit words), whereas the 803/503 were general purpose computers
> with 39-bit words.
>
> Elliotts were in the forefront of the whole concept of industrial
> automation: they probably sold more machines for that area than they did
> for GP work.
I believe there was something called a "Panelite 803" (Spelling ???)
Which was an 803 with extra interfaces for process control. I'm sure
I've heard that an 803 was used to control a steel rolling mill somewhere,
and I have an article somewhere from a GPO journal describing the 803
system that was used to prepare paper tapes that were used to steer the
early satellite dishes at the Goonhilly earth station.
Peter Onion.
>
simon_p...@hotmail.com
The Algol programs were on 8-track paper tape via 1000 char/sec readers.
I used the Elliott simulation package for an economic model of local cable provisioning.
At one time I had a printout of a complete memory dump and was able to identify my program routines and where all the variables and constants were stored. mnemonic codes 20, 30 and 24 are still etched in my brain!
Bob Eager
https://computerconservationsociety.org/resurrection.htm
--
Using UNIX since v6 (1975)...
Use the BIG mirror service in the UK:
http://www.mirrorservice.org
Eric vdM
The beauty of systems like the Elliott was that it was possible to understand all of the hardware and part of the software. And remember all 64 of its functions codes, in octal.
You may be interested in having a look at some of the manuals for the Elliott 503. They can be found at Bill Purvis's site: <http://www.billp.org/ccs/503/index.html>
Eric van der Meer
Quadibloc
> Reviving a twenty year old thread. Wow.
there had been another computer which was a renamed RCA machine.
Just now, I noticed that the Elliott 503 has another claim to fame; it was the
computer that used Cluff-Foster-Idelson code, which was the first computer code
to have a general arrangement of characters similar to what would later be used
in ASCII (although it was definitely very different from ASCII in a number of
ways).
John Savard