|
Preservation
1900 Hardware Emulators
1830 GP VDU
Operating Systems
Programming Languages
Software Tapes
Technical Publications
George 3 Executive Emulator
A Brief History of ICL
(new window)
ICL 1900 Pages
(new window)
GEORGE 3 Pages
(new window)
Home Page
|
|
The ICL 1900 series computer had a variety of operating
Systems available from ICL ranging from the basic Operators
Executive to the comprehensive George 3/4 systems, all of
which were provided free of charge as part of the system.
The only exceptions to this, as far as I remember, were the
'+' series from ICL Dataskil (ICL's software house).
| EXECutive |
|
Operators Executive was the basic
manual operating system for all machines, except the
1906A & 1906S. There were different versions for the
'Stevenage' and 'West Gorton' systems and it was
always generated specifically for a site to reflect
the actual machine configuration. Despite the
differing versions, they always presented the same
user interface.
Operators would load and run (multiple in most
versions) programs under its control, from
instructions provided on the 'job control' sheet.
Unlike IBM operating systems (from what I have
read), no partitions were defined, all resources were
available and allocated to programs on a dynamic
basis.
Current status: We have source tapes for E6RM and
EWG3 and the generation utilities and can successfully
generate both Executives, which are running on the
1904S Hardware Emulator.
In addition we have some complete and partial Executives
for some Stevenage machines, either as listings or in
binary format on cards/paper tape.
|
| Automatic Operator |
|
Automatic Operator was an enhancement to single
programming EXECutives, which allowed a series of
programs to be successively loaded and executed with
a minimum of operator intervention. The commands
being read from the Card/Paper Tape Reader; a very
basic form of Job Control.
Current status: This appears to be
lost.
|
| GEORGE 1 (GEneral ORGanizational Environment) |
|
George 1 was a batch operating system orginally
designed as a more sophisticated version of Automatic
Operator. It was initially implemented for "West Gorton"
executives using the "two-slot trusted system". This
produced the orginal George1T (with overlays on magnetic
tape), and used the same Job Description Language as
George 2. There is a common
'Operating System Format Reference
Card' for the two systems.
Stevenage EXECcutives were initially not multi-program
and could not run George 1 (never mind George 2). So
the "Single Slot Trusted System" otherwise known as the
Range Compatible Trusted Program system was developed and eventually
replaced "double-slot trusted" even on West Gorton
EXECs. Using SSTS, George 1 was re-implemented for all
machines from 1901 up to 1905 and subsequently all
1900s.
Current status: We have both the MT and ED overlaid
versions on an issue tape, but have made no attempt to run
them so far.
|
| GEORGE 1S |
|
George 1S was produced by Jo Morritt together with the
Stevenage executive team for the 1901A. The basic
control sequence of Job Analyser, Runjob, Endjob was
hard-coded into the overlaid exec and Jo re-wrote Job
Analyser as a subject program rather than an overlay
(with exactly the same JDL).
Current status: This appears to be lost.
|
| GEORGE 2 |
|
George 2 was a batch operating system designed for
small/medium machines, and was George 1 with spooling
(for basic peripherals) and a 'job well' so that jobs
could be read in by an Input Module ready for a Central
Module to become available. The output created during
the run was printed (or punched) by an Output Module.
The system ran as a series of programs loaded under
Operators EXECutive. Typically, one Input Module was
loaded, one Output Module per line printer and a number
of Central Modules depending on the processor power,
available memory and other resources available. The
Central Modules actually controlled the jobs,
interpreting the JDL, loading and running the
required programs. There were different variants of
George 2:-
- G2EE - disc overlay/disc spooling
- G2TT - tape overlay/tape spooling
- G2ET - disc overlay/tape spooling
- G2TE - tape overlay/disc spooling
all of which were produced from a common source tape
(including G1T and G1E). The total source code including
the Loader, Job Analyser, RunJob, Endjob, Dump, Print
for both disc and tape variants ran to about 30,000
lines of GIN.
Current status: We have issue tapes for mk.9G and mk.9H and
some source tapes. The disc version (G2EE/9G), with input and output
spoolers, loaded from an issue tape is happily running under E6RM
on the 1904S Hardware Emulator.
|
| GEORGE 2+ |
|
George 2+ was an enhanced version of George 2
provided by ICL Dataskil. Some of the enhancements
included additional JCL facilities, a job scheduler
and a spooler for the LPS14 Laser Printer (mea
culper).
Current status: Sources and binaries for various
GEORGE 2+ versions have been recovered.
|
| GEORGE 3 |
|
George 3 was the flagship
operating system designed for medium/large machines.
It provided both batch and MOP (multiple online
programming) services, together with a Filestore
(similar to a Unix file system) and a powerful Macro
(JCL) language. Also incorporated was an incremental
backup system (Dumper) for the Filestore and a tape
librarian.
Under George 3, basic peripherals (card
reader/punch, paper tape reader/punch and line
printers) were rarely used online, the input and
output being off-lined via the Filestore. It was also
possible to simulate small magnetic tape and disc
files within the Filestore to avoid using real tapes
and discs for small amounts of data.
The size of the Filestore was not dependant on the
disc storage allocated to it, this provided the
'online' Filestore. As long as sufficient safe copies
of a file were held by the Dumper system, it could be
deleted from the 'online' Filestore and was then said
to be 'offline'. The user had to do nothing when the
file was required, it would be automatically
retrieved from a Dump Tape when an attempt was made
to use it. To speed jobs up, a list of files required
could be specified at the beginning of a job (or as a
seperate preceeding job) requesting that the required
files be brought back 'online'.
It had its own special Executive, which was limited
in function (much more like the BIOS on a PC) as all
of the work was run via George 3, apart from the initial
system load.
Current status: Working binaries, partial (user
issue) source and software notices are available for Mk.
8.67. Some items from earlier versions have also been
recovered. George 3 mk.8.67 is happily
running both on the George 3 Executive
Emulator and under EWG3 on the 1904S
Hardware Emulator.
|
| GEORGE 3+ |
|
George 3+ was not an operating system, but a set of
enhancements/additions to the standard
George 3 system provided by ICL
Dataskil. They were also applicable to George 4
systems.
Current status: Some GEORGE 3+ enhancments have
been incorporated into GEORGE 3 8.67. The LPS14
issue tape has been recovered.
|
| GEORGE 4 |
|
George 4 was the same as George 3, except that it was
designed to run on the paged memory machines (1906A &
1906S).
Current status: This appears to be lost, an
earlier version of the GEORGE 3 full source tapes
would be needed to recreate GEORGE 4. In addition,
no paged utilities appear to have survived.
|
| MINIMOP |
|
Minimop was a free-standing small Multiple Online
Programming system, that could run alongside George 2
assuming that the system had sufficient capacity,
providing access to the machine from terminals (both
local and reomte), usually teletypes.
Current status: This appears to be
lost.
|
| MAXIMOP |
|
Maximop was a much improved Multiple Online
Programming system than Minimop, was written by
Queen Mary College, University of London, and was
distributed by ICL as a replacement for MINIMOP.
Again, it could run alongside George 2 assuming that
the system had sufficient capacity.
Current status: An originnal issue tape has been
recovered and Maximop/Maxibatch is happily running under
E6RM on the 1904S Hardware Emulator.
|
| PATSY (Programmer's Automatic Testing SYstem) |
|
PATSY was a complete program amending, compiling
and testing routine designed for use with single
program EXECutives for programs written in PLAN,
COBOL or NICOL, and held on a magnetic tape in source
form. The testing phase loads the progams and
operates them according to the operating directives
provided by the user and held on the source file with
the program. See manual TP4018 for more details.
Current status: This appears to be
lost.
|
Combined George 2(+)/Maximop was quite a common medium-range
(1902S/1903A) configuration. It was also used on larger
machines (1904A/1904S), where sites didn't want the
overheads of George 3 which gave you a very sophisticated
system, but at a price in performance.
In addition to the Operating Systems provided by ICL, some
sites (generally universities or similar educational
establishments) chose to write their own Operating System,
presumably streamlined to suit their particular workload,
rather than use one of the general purposes systems
provided.
| Eldon3 |
University of Leeds
ca.1976 |
A Terminal/MOP system that ran on a
1906A under GEORGE 4.
|
| FAAST |
University of Leeds
ca.1976 |
Fortran, Algol and Algol68 Short Turnround batch
system that ran on a 1906A under GEORGE 4.
|
| OPED |
Defence ADP Training Centre (DADPTC Blandford)
ca.1968-69 |
When GEORGE 2 was available at pre-release, there was
a requirement for an O/S suitable for a teaching
environment. The result was OPED, a basic batch O/S
designed to run on a 1903 and written by Dr. Kenneth
Hunt and Major (now Professor) Tony Sammes.
|
| SUMP |
RMCS Shrivenham
ca.1973-74 |
The Shrivenham Universal Management Program was a
batch O/S especially well-suited to pushing
through lots of small students' jobs quickly; it was
written by David Sugden and John Hunter as a rewrite
of OPED, designed to run on the new 1903S. The system
was restricted to running up to 3 batch streams and
interfacing with MAXIMOP. SUMP provided:
- A file system with rudimentary access controls
- A very flexible JCL (with lots of bells &
whistles at the request of academic customers)
- The ability to save a part run core image
and resume it when SUMP was next run.
- Halving of the Computer Room staff requirement
(having fired up the system in the morning, all
they had to do was change magnetic tapes, feed in
cards and paper tapes, feed the lineprinters and
distribute the outputs into pidgeon holes.
The system lasted until about 1981, when the college
changed over to a VAX.
|
| TPOS |
ICL |
Test Program Operating System was developed for
internal use. Its purpose was to run a cycle of
system acceptance tests, collecting evidence of
faults, for unattended EXECutive and System
testing
|
|