EDSAC — Andrew Herbert

We continue to make good forward progress with EDSAC and continue to enjoy an extended period with very few component failures.

Since the January report the most significant step forward has been to demonstrate writing of the results of computations in the Arithmetic Unit being successfully written back to store (in the right location and with the correct value) via the Transfer Unit. This required some work on the interface between the Arithmetic Unit and the “Master Output Bus” and between there and the Transfer Unit. The MOB is where several data sources are merged for machine functions that write to store (data input via I order, arithmetic results via T and U orders, initial orders and engineers’ switches). Each source has to be adjusted to bring signal levels into line, and timing adjusted to ensure each is brought back into synchronisation with the main store, whose timing effectively controls the whole machine.

Work on the paper tape reader has been held up while the Transfer Unit is being checked out. However, in initial testing we found a design fault in the timing of the logic to drive physical reader and a small redesign is in progress.

We have overcome most of the problems previously experienced with the operator control buttons by adjusting signal levels, although in recent weeks the (store) CLEAR button appears to be toggling the machine “stopped/running flip-flop” but this is mostly harmless and will be investigated in due course.

We have also overcome the previously reported problems with running short programs, accessing operands with addresses close to the current order, and single-stepping through programs. Many of these problems were eventually attributed to the predominance of AC coupling in the EDSAC design and the consequent build up (or leakage) of charge in circuits holding flip-flops on or off and making their behaviour sensitive to the timing and sequence of orders being executed.

The final new development is the addition of main store monitoring via our network of microprocessor based “EDLA (EDSAC Distributed Logic Analyser)” probes around the machine. These capture data from EDSAC when running and send them back via WiFi to a central server for subsequent analysis. This system can now show us the behaviour of the clock and digit pulse system and also reads from, and writes to, the store. The latter is a particularly useful diagnostic tool since we can confirm the contents of the store before and after a test program is run. The system also maintains statistics on signal levels and their shaping, giving us a useful maintenance aid in spotting components that may be drifting out of the desired margins.

Recent focus has been on the complementor in the arithmetic unit. Adjustments have been made to internal delay lines to correct timing of search for first non-zero bit and gating to produce correct output.

Some changes have been made to -50V DC power distribution to reduce excessive leakage currents.

There is now an improved interface between Signal Sequencer Injector (SSI) and main store addressing logic to improve reliability of SSI as means to load test programs.

Main store rack F2 is is being re-enabled after SSI improvements.

Schematics for main store addressing were updated to be consistent with relevant chassis and capture recent modifications.

The much-hacked chassis 27 in Main Control is being reworked to make modifications permanent.

Investigation of scope for improving quality of register CRTs in display units is being undertaken.

We are reconstructing the tape reader control chassis – replacement of timing monostable by phantastron.

ICL 2966 — Delwyn Holroyd

The store cabinet air flow sensor developed a fault that caused the sensor to trigger shortly after first power up, but once things had warmed up a bit it would be OK for the rest of the day. The airflow sensor consists of a device containing a heater and a bi-metallic strip installed in a valve socket. If insufficient air passes through then the heat causes the bi-metallic strip to bend and break the circuit. The sensor is positioned directly underneath the card platters and behind the power supplies, making it extremely difficult to get to. After taking it apart and cleaning, it seems to be working again although I suspect a bad contact in the valve socket was the real issue.

Harwell Dekatron — Delwyn Holroyd

The machine has suffered a succession of issues caused by dirty relay contacts in recent months. One particularly amusing fault caused arithmetic subtract orders to produce completely incorrect results! This was traced to a dirty contact propagating the carry signal between digits. The problem had been noticed in the context of division, which involves repeated subtraction and shifting.

There are still at least two other problems with division – one where the machine stops during the operation and a second where it starts giving an incorrect answer but only after some hours of running.

Another current problem is that the machine will stop with multiple locations in one store selected, which ought not to be possible! This is proving very tricky to diagnose because it happens only rarely and once in this state it’s difficult to see how it came about.

IBM 360 — Adam Bradley

It is with much regret that the team who rescued two complete IBM 360/20 systems and a 370/125 have concluded that a combination of rising accommodation costs and insufficient time away from their proper jobs is incompatible with their objective of bringing the 360s back to life. All is not lost, however, as a new, very suitable home has been found for the machines for the next 20 years, where they will be publicly displayed. After this period, the machines will be returned to the original team, who by then expect to be rather more available.

The Argus 700 has enjoyed a good period of serviceability and is ready to be demonstrated to the public at its new and permanent home, the RAF Air Defence Radar Museum at Neatishead.

Current work includes checking the serviceability and condition of spares plus labelling and boxing. Other tasks include investigation and development of bench testing various A700 PCBs. Many Argus 700 items recovered from Switzerland have yet to be checked.

Documentation is being created for knowledge transfer, including fault finding techniques, disc formatting etc. – a means of consolidating ten years of learning and experience.

Our Argus 700 based engagement simulator has featured in the introduction for a Sky News item shown a couple of weeks ago. If you are interested the item is available on YouTube at wwww.ccsoc.org/argus0.htm. The item is about the UK’s ability, or lack of it, to defend itself as it did during the Cold War. Note: I do not figure in the filming.

Elliott 803, 903 & 920M — Terry Froggatt

Many CCS members will know John Sinclair, who restored both the Science Museum’s 803 and TNMoC’s 803. He handed the maintenance of TNMoC’s 803 over to Peter Onion some years ago, but we were pleased to see him in the museum again when he attended the TNMoC volunteers’ “late Christmas” party early in February. Of course, on that day the 803 developed a store fault, which Peter says gave him the chance to get his “803 Engineer” certificate re-validated by a real Elliott Field Service Engineer. The problem was a faulty SZ2 element. These drive the store plane inhibit wires and stop the writing of ones which would otherwise happen at the end of every word time. The SZ2 is triggered when a zero is to be written. This leaves the core in the zero state established by the previous read operation at the start of the word time.

Peter Williamson reported that the TNMOC 903 was working early in the year, and I can add that it was still running BASIC demonstration programs , when I visited TNMoC at Easter.

As reported in res104.htm#terry, somewhat to my surprise, I’ve got the RAA/TNMoC 920M (88) working again, and I’ve subsequently had its audio output monitor working too. I’ve now made my analysis of the 920M workings available at www.tjfroggatt.plus.com/MCM7.HTM, although I would caution anyone thinking of buying a 920M that the store details which I’ve given are far from complete, and that this is where the problems usually are.

But also as reported in res104.htm#terry, somewhat to my disappointment, I’ve been unable to repair the eBay 920M (5343). There is problem with the type 38 module (at position L04), probably a leaking capacitor between the 0v & 12v pins. Erik Baigar’s spare 38 module also shorted as soon as I applied power to it. We have been able to work round capacitor shorts in other modules, by isolating them, but this only works when at least one end of the capacitor is not connected elsewhere internally.

Also when I ran 5343 with the type 38 borrowed from 88, it would run some programs, but one word in 64 was failing, suggesting a store diode failure. But I was unable to locate the fault, and there is a 50% chance that a failed diode is inaccessible. So with at least two difficult faults in 5343, I decided it was time to move on. I refitted 88’s borrowed type 38 module, and refitted its covers. It is now ready to be returned to TNMoC, where hopefully it can be put on display, to show its unique construction.

Whilst testing 920M (88) with as many suitable programs as I had to hand, I discovered a bug in my own BCPL implementation, in which the B-register is frequently used as the stack pointer. The normal 900-series subroutine exit also uses the B-register, so to avoid having to reload the stack pointer after subroutine calls, I’ve used unusual exit code, which adds the link to an unmodified jump instruction then executes it. Alone in the 900-series, and as clearly stated in its Facts Cards, the 920M puts junk into the two top bits of the subroutine link. I’d overlooked the need for a mask instruction to remove this junk, which would add just one word to the exit code shared by all subroutines.

IBM Museum — Peter Short

Curators have been focussing on two main activities; trying to bring more hardware that is on display back to life, and sorting out any excess machines (i.e. where we have more than two) for disposal. We are also looking at potential new display items and potential rearrangements within the museum.

Hardware Repair

Typewriters

We have successfully repaired one each of the three typewriter technologies:

• Typebar, using a model B
• Selectric
• Wheelwriter

These have been moved to our ‘touchy feely corner’ so visitors can see hands-on the technical progress of this now outdated technology, and also try them out. This should also make it easier to explain to the younger generations what a typewriter is. The Selectric input/output typewriter from a 2741 terminal has been on the workbench for years. We thought the job was finally done, only to find that the keyboard keys were all slow to restore. The culprit has been identified; each key lever has a small pivoting spring-loaded bracket that pivots out of the way when a key is pressed, relieving the load on the return leaf spring. Every one of these is jammed solid, and can only be freed up by a combination of cleaning fluid and physically moving the bracket, which is virtually impossible in situ. We are now in the middle of extracting the key levers and freeing them up. This could take some time!

DisplayWriter

DisplayWriter was a dedicated word processing computer, released in June 1980, a little over a year before the IBM PC was announced. We have had one on display for many years, and previous attempts to get it working failed. Some time ago a 110V DisplayWriter system was donated, and using the system unit from this machine has enabled us to bring it back to life, The only 8” diskette we have is a diagnostic disc, so if by chance anyone has some word processing software we could copy we’d be most grateful.

Maintenance Device

Maintenance Device, or MD, was a portable diagnostic machine for the 3380 disk array and 3880 controller. Over the years we have collected quite a few of these. The only one that we know is working is on display, but without a 3380 there’s not much more we can do. We’re sorting through the remaining MDs so we can keep one of each of the two versions in the collection.

Personal Computer

Our PC-XT has been running constantly for quite a while, usually with the IBM Alley Cat game loaded. It failed a couple of weeks ago and has been replaced with a working spare for now until we get time to see if it’s repairable.

IBM and NASA

Following Dik’s appeal on behalf of the BCS magazine editor for help for an article about Computers in Space, we are now in contact with Martin Cooper. IBM’s relationship with NASA goes back to the very earliest times in the 1950s and has continued to this day. It’s difficult to see how the huge efforts can be netted down to a magazine article; perhaps we will need to concentrate on one programme, or take over a whole issue!

For example the Moon landing was supported by around 4,000 IBMers who built the computers, hardware and software to guide and track Apollo from liftoff to splashdown, and provided ground support with five S/360 Model 75s.

The Saturn V Instrument Unit (IU), the top black ring on the final stage of the rocket, took over control from lift-off through Earth orbit to setting the correct course to the Moon. It managed everything from telemetry, speed, velocity and thrust to guidance, radio communications and emergency detection systems. The 2 ton, 21 feet (6.4M) diameter by 3 feet (0.9M) high unit, lined with mission critical instruments, was designed and built by IBM. It had a MTBF of 40,000 hours and a component density of 250,000 per cubic foot or nearly 9 million per cubic metre, if my maths still works properly.

The new IBM Heritage area of the IBM website has several pages covering the space programmes. The Apollo page is www.ibm.com/history/apollo where you’ll find links to all the other space programmes and many other interesting pages on other IBM innovations and the company heritage via the “Advancing Humanity” tab.

None of this involved Hursley Museum, although one of our curators was involved with the Shuttle programme.

The National Museum of Computing — Aneesa Riffat

Having submitted our application to the Arts Council to become a Nationally Styled Museum through the Accreditation Scheme, we are due a site visit shortly.

We are currently reviewing items that are on loan to and from the Museum. All loans will be fixed with a term, which can be renewed if necessary.

Donations to the Museum are steady. There is a big variation in what is being offered, but we are pleased to report we have been acquiring more social history items into the collection in order to diversify our storytelling.

Watford Electric Totalizer

Donations of note include:

• A charming Lego Cray model made by an enthusiast.
• A Watford Electric Totalizer which was used by the Daily Mail in Fleet Street in the 1930s/40s. The donor is working with us on providing provenance for the item. It is in remarkably good condition given its age. We see this as a precursor to computing.
• A test fixture for fault finding on Ferranti Argus’ 500 CPU and core memory modules. We are led to believe that this is the last one in the world! We are currently working on repairing this item.

As we have been taking stock of what is in our collection, we expect to shortly begin a rationalisation project alongside inventory. We have amassed a large number of items in certain areas of the collection and would like to reduce.

Partnerships with Reading Museum in collaboration with DEXODUS and DECUS are continuing for us to exhibit and celebrate the 60^th anniversary of DEC. The exhibition will go live at TNMoC in the summer before moving to Reading in early 2025.

Bletchley Park Trust — Erica Monro

Much has happened at Bletchley Park recently, including:

• the opening of the Collections Centre which contains top-of-the-range environment controls and fire-suppression system in the storage areas, a dedicated digitisation suite and the Christopher Moore Reading Room for researchers;
• the opening of our Block E Learning Centre, housing nine classrooms equipped with interactive Lynx whiteboards running Clevertouch software;
• the opening of the Fellowship Auditorium – a modern 250-seat theatre, supported by a dedicated events and catering space (the Radley Room);
• the global AI Safety Summit in November 2023 which brought world leaders and industry heads to Bletchley Park to discuss and debate AI regulation, safety and development.

Research and content development is just beginning on our next temporary exhibition which will focus on artificial intelligence and open in early 2025.

At the moment, we are busy preparing for the 80^th anniversary of D-Day with activities and events on site including:

• Our major exhibition D-Day: Interception, Intelligence, Invasion;
• A one-day symposium on 1^st June, ‘Operation Overlord: The Intelligence Story behind D-Day’, with speakers including Steve Zaloga;
• A pop-up display of illustrations about the 1944 Hut 8 temporary Y Station, opening 16^th May (in Hut 8);
• A QR-code ‘D-Day spotlights’ tour of outdoor areas of the BP site;
• Family activities and D-Day themed family tours during May half term.

We have also begun a programme of monthly public lectures. One of our first talks was an excellent panel talk in January to mark Colossus’ 80^th anniversary, featuring David Abrutat of GCHQ, Phil Hayes of TNMoC and Martin Gillow (virtualcolossus.co.uk/).

Planning is underway for the inaugural National Intelligence History Conference, hosted by Bletchley Park and GCHQ, which will be held in the Fellowship Auditorium, on 20-22^nd November 2024.

Software Preservation — David Holdsworth

Atlas Emulation

Dik Leatherdale reports that after a long pause he has recently added some new functionality to his Atlas 1 emulator. The emulator now supports reading magnetic tapes written by the Ferranti Orion. Or rather, it supports a format which simulates Orion tapes. Note, however that Atlas did not support the writing of Orion tapes. Accordingly, in order to provide test data, a “blister” has been added to the emulator to allow simulated Orion tapes to be generated.

Elliott Algol

Andrew Herbert has implemented an integrated preservation of documentation and source code for Elliott Algol. Whereas our KDF9 Kidsgrove Algol equivalent and Leo III equivalent start from the source code and link to the documentation, the Elliott Algol implementation comes with documentation written by the authors and Andrew’s implementation starts from the documentation and links in the source code. It shows real promise in providing understanding of source code that is almost devoid of comments and is written in an assembly language (SIR) with numeric instruction codes. I have ideas for further embellishment.

The restrictions placed on the Algol acceptable to this implementation and the examples of implementation led me to think about a couple of Algol 60 features that have withered on the vine, viz call-by-name and switches. With regard to switches and gotos it is surely a matter of good riddance. I used to be a fan of Jensen’s device which was an elegant exploitation of call-by-name that may not have been foreseen by the designers of Algol 60. Looking at the implementations of non-local variable access (particularly tangled in call-by-name) I can see why these features got discarded as difficult to implement, but more modern compilers are cleverer.

Jensen’s Device Here is a reminder of one of Algol 60’s more useful quirks. Perhaps I am still a fan.

A procedure with the heading:

        'real' 'procedure' int(a, b, x, f);
        'comment' integrates f(x) over range a to b;
        'comment' x and f are both called by name;
        'value' a, b; 'real' a, b, x, f;

can be called:

        int(0, halfPi, th * sin(th));

to integrate:

∫₀^π/2 Θ sin Θ dΘ

The same computation in C or Java involves multiple procedures. If you try to program a modern equivalent of:

        int(0, 1, x, int(0, x, y, f(x,y)));

it gets very hairy. It all works on KDF9 Algol.

News Round-Up

Herbert Bruderer has a list of abaci in museums at www.ccsoc.org/bru15.htm.

101010101

This edition of Resurrection must sadly report the passing of six individuals who may be well known to members of the Computer Conservation Society.

Peter Short of the IBM Museum at Hursley Park passed away not long after submitting his last report for Resurrection which readers will find above. His obituary by Peter Coghlan follows below.

Aneesa Riffatt was a senior Curator and Collections Manager at The National Museum of Computing and was responsible for liaison with us at the Computer Conservation Society. Although she had not been in post for very long, it was evident that her enthusiasm and dedication was a huge asset. Ever helpful and ever quick to respond, her death was unexpected. We had exchanged emails on the day of her passing. Her last TNMoC activity report is also above.

Our good friend Brian Spoor, a member of a small group of CCS experts on the ICT/ICL 1900 Series and a frequent contributor to these pages passed away in May. An obituary is planned for the next issue of Resurrection.

Bernard Richards, Turing’s last PhD student, has also died, severing a significant link with the past. He was a Professor of Medical Informatics at Manchester University and he spent much time working in the area of medicine through his computer systems. Later, he worked on expert systems aimed at use in open heart surgery and also intensive care units. He was a member of the CCS North West Group.

Finally, Roger Johnson reports the passing of both Prof. Ross Anderson FRS of Cambridge and Edinburgh Universities, a veteran of many important technical and policy campaigns, and Gary Tee, among many things, a DEUCE programmer and New Zealand’s first historian of computing.

May they all rest in peace.

From time to time, CCS officers receive requests for historic computer information from CCS members and from the general public. Our usual practice is to circulate such queries to the CCS Committee whose members have, between them, an impressively wide knowledge of such matters. A recent query concerning the Canadian FERUT and UTEC computers was received. Within a couple of hours no fewer than three relevant responses had been sent. Sometimes we exceed expectations!

One of only five extant replicas of Doug Englebart’s original mouse of the mid-1960s.

The annual CCS European museum visit in April was to the Musée Bolo, the Swiss Museum of Computer Science, Digital Culture and Video Games www.museebolo.ch/en/, situated on the university campus by Lac Léman (Lake Geneva), the Ecole Polytechnique Fédérale de Lausanne (EPFL)

The museum was founded in 2002, the first of its kind in Switzerland. In 2007 the “Mémoires Informatiques” Foundation was established to manage the museum, in order to conserve and enhance IT heritage. Friends of the Musée Bolo, all volunteers, support the maintenance, preservation and promotion of the museum’s machines, material and documentation.

Cédric Gaudin, the President, was most informative in leading our guided tour, and answering all our questions.

The artefacts displayed at the museum comprise only three per cent of the collection of several thousand computers, peripherals, accessories, software and publications of various kinds. The large reserve collection is located in a storage facility of 100 square metres in central Lausanne.

CRAY 2 supercomputers and IBM Blue Gene (we had helpfully been circulated beforehand with a 155-page Briefing Pack on Cray History) are on display, as well as numbers of more familiar and smaller machines, such as the Commodore PET, DEC’s PDP-11 and Sinclair Research’s ZX-81. There is a significant collection of games and video consoles.

Those were the Crays, my friend. We thought they’d never end.

Naturally the permanent exhibition highlights a unique collection from the Swiss computer industry itself, which flourished in the early days of the microcomputer boom.

Our late afternoon visit was followed by an enjoyable meal at Gina Restaurante, also at the EPFL site.

Most of our party had arrived at the museum by bus or Metro from hotels in central Lausanne (visitors to the city are offered very convenient free travel passes), and we returned by Metro. But it was not the end of the evening for those of us who took advantage of Cédric’s invitation to view the storage facility. We learned more from him about the several thousand computers and other objects linked to IT history there. The museum is planning to increase its storage space to accommodate new acquisitions.

Apart from the main purpose of our expedition, during our stay we were able to explore Lausanne: the Old Town with its cathedral and other historic buildings, the other museums, and Lausanne’s lakeside gardens, port and views.

Several of us took a trip on Lac Léman by paddle ship, a steam powered vessel sailing up to Montreux, past the spectacular UNESCO World Heritage site of Lavaux Vineyards, calling at Vevey and Montreux and the Chateau de Chillon, and returning by train to Lausanne.

CERN was not too far away if flying to Geneva, and a few of our party took the opportunity of taking an individual tour, recommending it unreservedly: “Fascinating, well explained, friendly and all free.”

Another successful CCS expedition! We are most grateful to Cédric Gaudin for his welcome and for showing us round so knowledgably.

Many thanks to our organiser, Dan Hayton and to local CCS member Rob Watts who was a valuable source of useful information.

Programming the Later EDSAC

In Resurrection 103, Andrew Herbert, project lead for the EDSAC replica, published an interesting article about further developments of the basic EDSAC 1, commenting that all the EDSAC emulators about which he knows support that basic version. Andrew subsequently told me that, as far as he knew, there were no emulators around for the later version of EDSAC 1. As the author of one of the aforementioned emulators for EDSAC 1, this seemed like an interesting challenge.

Summary of the differences between EDSAC 1, and the “later” version.

The primary difference is the addition of a Modifier or Index Register, the B-Register, which can be used to modify the address field of any order for which the programmer might choose to involve the said register. This register can be used by any order which needs it, by having a bit, unused in the basic EDSAC, in the order set to 1. I refer to this bit as the “B-bit”.

R30 , listed below, which was written to enhance the initial orders’ capabilities, can be used to make life easier for the programmer.

For example. if you assume that the B-register contains the value 10, the order T 100 F will save the contents of the accumulator to location 100. To invoke the use of the B-register, it would be possible to use the order T 1124 F (where 1124 = 100 + 1024). If subroutine R30 has been loaded, it becomes possible to use the order TS 100 F as a possibly simpler alternative. In either case, the destination address is modified to 100 plus the contents of the B-register (i.e., 110).

The following table, based on Andrew’s article, lists, in alphabetical order, the new/altered orders which have to be implemented to support this new functionality.

* Note that orders in the table above marked with an “*” are required to allow support of the R30 subroutine mentioned above.

^{1. Of course, the Colossus actually used physical 5-track tapes, but there are two “virtual” tracks to indicate “start of message” (SOM) and “end of message” (EOM), and ‘my’ K function allowed appropriately written software to read simulated Colossus tapes including these SOM and EOM bits.}

Upgrading my emulator.

The changes as specified above were duly made to my emulator, which was a reasonably simple process, except for one “gotcha” found during testing. This concerned the Left and Right Shift orders (“L” and “R”). These two orders have the interesting quirk that they use the 10 address bits and the “spare” bit to the left thereof to specify the number of places to be shifted ². This means that my system initially triggered the use of the B-register at quite inappropriate moments causing a certain amount of head-scratching.

Once that was resolved, testing could then continue.

Firstly, a program was written to create the squares of all integers in the range 1 to 255; the program was designed to save these squares in sequential memory using the newly implemented B order. This is possibly the first program to have been written for the later version of EDSAC 1 for some 60 years! Its listing is below:

To explain further, the table of squares produced by this program contains all squares from 1^2 to 255^2; this upper limit was chosen as larger numbers would involve double length arithmetic, which is perfectly feasible, but not the point of this exercise. Given that the selected base of the table of squares would be 150 , this would place the upper end of the table at 404. This means that the process should stop if the B-register hits 405.

^{2. The exact operation of the shift instructions is expanded upon in Appendix I.}

To start with, the B-register is initialised by the instruction at address 100 (B 150 F) to point at the base of the table.

Then, after a particular square number has been calculated, it is saved into the list of the squares using the instruction US 0 F. This is the “modified” version of the U order, where the modification (that is, the use of the B-register, is indicated by the extra S immediately following the U order. The B-register is then incremented, ready for the next loop, using the self-referencing instruction BS 1 F at address 108.

Finally, a decision is made as to whether to continue around the loop by subtracting 405 from the B-register using BS 405 S; the decision is made by the new J order which jumps if the result of this subtraction is non-zero, and drops through if the result is zero to terminate at the Z F. If the result is non-zero, it is necessary to add the value 405 back into the B-register using BS 405 F.

Below is a fragment of a store dump³, where the memory was written using the above program using the ‘new’ B-register as an index.

I then thought that it would be interesting to print the table of squares; my emulator, in common with others, has a print window, and also the capability to print to an external printer. There exists a standard subroutine for printing short integers in the EDSAC library for EDSAC 1, known as P6; this subroutine prints a fixed-format 5-digit decimal number, with leading zeroes replaced by spaces. I decided to write a replacement for this later version of EDSAC, to be called P6B.

In the updated manual for the later version of EDSAC, it is pointed out that subroutines have, generally speaking, to be rewritten to take advantage of the enhanced/changed facilities, but that turned out to be rather more difficult than originally expected.

The problem in this particular subroutine revolved around the F order, which in the original EDSAC, was sued to read back the relays in the printer so that it could be checked that the printer is really printing the character which was sent to it. As noted above, this function has been repurposed to save the contents of the B-register in the form of an executable order for later recovery of the B register’s contents.

^{3.First and last six lines only}

After in-depth study, it turned out that the P6 subroutine uses the F order, not for checking the printer’s relays, but as part of the algorithm converting the binary format of the number to be printed into printable decimal. Replacing this application of the F order required a fair amount of extra coding.⁴ Such coding, for reasons which aren’t important right now, involved the use of the C (Collate) order and the Multiplier register which the calling main programme was using. There is no order to save the Multiplier register, so its content has to be used as part of a multiplication by unity, resulting in a copy of the Multiplier register appearing in the Accumulator, which itself can be saved.

The new B order somewhat simplifies the call-and-return procedure for subroutines and it is fairly simple to preserve the current value of the B-register before using it for the call-and-return procedure. The B-register’s contents can be saved using the new K order into an instruction in the yet-to-be-executed instruction sequence.

The above calling sequence works as follows; all instruction addresses are shown in the source above in square brackets [thus]:

• In the main program, a K order at 112 saves an instruction to recover the B-register to 115, overwriting the dummy P order there.
• The B-register is then loaded with the address of instruction 113 using a B order.
• The following F order then makes an unconditional jump to address 500 which in this case is the first instruction of the P6B subroutine.

The subroutine itself then saves a recover-the-B-register instruction to 52Θ⁵ to enable that subroutine to use that register for its own purposes.

Once the subroutine has completed its work, it recovers the B-register’s contents and uses the instruction FS 2 F jump back to address 115 in the main program.

^{4. The detail of this is that the digit to be printed appears as five bits at the most significant end of a 17-bit word, and the following 12 bits contain the rest of the number to be processed later by the calling program; the output order (O) and/or the printer ignore the trailing 12 bits and only use the leading 5 bits.
The F order, on the other hand, returns the five bits from the relays, with 12 zero bits following. The difference is important, and had to be taken into account of in the replacement subroutine.}

^{5. The ‘Θ’ character here indicates that the address is relative to the subroutine‘s base address.}

This image shows the output from the printer, as logged to a text file.

A note about the “Greek”.

On reading the available plethora of documentation concerning EDSAC, it could be seen that various Greek letters were used as part of the programming character set. These are (upper case) Θ, Φ, Δ and (lower case) π for the various emulators which I’ve seen including my own, these characters are typed as @, !, & and #, respectively. As it seemed likely that I would need to retype various programs and subroutines to test my revised emulator, the easiest route seemed to be to add support for the Greek characters; however, to avoid having to add support for UTF-8 or Unicode, the simplest solution was to maintain 8-bit only input by using code page 1253, along with an appropriate font, for data entry. Upper case Σ also makes an appearance as a printable character.

From Wikipedia

Acknowledgements.

My thanks to Andrew Herbert for his recent article in Resurrection 103, which introduced me to this new “rabbit hole”, and who subsequently gave me some useful words of advice.

Appendix I - the L & R shift orders.

As noted above, the general layout of instructions on the later EDSAC 1 is thus:

L & R L & R shift orders use the position of the rightmost bit which has the value 1.

Therefore, if the Long/Short bit is set to 1, then there is a shift of one to the left or to the right. On the other hand, if the Long/Short bit is set to 0, then the shift is determined by the right-most bit in the 10 bits comprising the address field. If the address is 0, but the B-bit, which is a spare bit in the earlier EDSAC, is set to 1, then that is used to determine the shift count. If that value is also 0, then the right-most bit of the order is used.

As it happens, the right-most bit with value 1 of the L order is in a different position to that of the R order, The full sequence is thus:

^{7. Also note that the value1024 has overflowed into the B-bit (which was spare in the earlier EDSAC). The use of the B-register is suppressed in these two orders.}

^{8. This happens as the bit pattern for an “ L” order is 11001.}

^{8. This happens as the bit pattern for an “ L” order is 11001.}

Tony Sale Award 2024

Fabien Wernli representing Association MO5 accepted the Award and gave a splendid history of the Micral N and of the project to reconstruct it.

The Tony Sale Award began in 2012 as a biennial event. Because of the interruption of the pandemic, the 2024 Award was the first since 2018. The Award has always been international in its scope, and this year was no exception. We had a total of five entries, one from France, two from the United States, one from Ukraine, and one from the UK.

In keeping with the international scope of the Award, there was an internationally representative jury. The judges included Paul Ceruzzi, curator emeritus at the National Air and Space Museum, Washington D.C., where he oversaw the computing exhibits. We were delighted to have David Link, an artist and media theorist from Germany – who also won the 2012 Tony Sale Award. And from the UK we had Professor Simon Lavington, a distinguished author of several books on early British computing. From Wales we had Professor John Tucker, formerly pro-vice chancellor of the University of Swansea and warden of the University’s History of Computing Collection. And lastly, we had Professor Jonathan Bowen, chair of the BCS Specialist Group on Formal Aspects of Computing Science and an authority on Alan Turing.

All the judging panel were impressed by the very high standard of the submissions for the Award. Every one of the entrants was a potential winner, and worthy of an Award. Interestingly, this year three of the submissions were for “virtual” rather than physical reconstructions of early computers: there was a web-based virtualization of the ICT 1301 computer from the UK; from Ukraine there was a simulation and software reconstruction of the country’s second digital computer, the Kyiv; and from the United States the SIMH project enables the emulation of practically any digital computer. A fourth project was the restoration of Medley Interlisp, Xerox PARC’s version of the Lisp programming language initiated in the 1960s.

However, there could only be one winner. After a great deal of deliberation by the jury, the Tony Sale Award for 2024 went to France’s Association MO5 for the hardware and software reconstruction of the Micral N microcomputer of 1972. This was not only France’s first microcomputer, but among the very first in the world.

Obituary : Peter Short

Peter first volunteered as a museum curator in 2011 and quickly realised that the museum needed some external visibility, and so started work on the museum’s website.

Over the years the website has continued to evolve into a comprehensive exhibit catalogue and the far more extensive portal that you see today, all of which was designed, built and managed by Peter.

Peter was passionate about preserving IBM’s historical legacy and was heavily involved with the IBM Corporate Archive, other IBM museums in Europe and further afield; computer preservation and conservation societies and individuals, including the British Computer Society, Computer Conservation Society, Bletchley Park and many more.

Peter was a beacon of diligence, tirelessly working to document, protect, and share the stories of IBM’s computing history.

Along with his colleagues, Peter created exhibits that brought the evolution of technology to life, inspiring countless visitors to marvel at the pace and complexity of innovation or to relive their own experiences in the fledgling early days of computing.

Peter’s passing leaves the IBM Hursley Museum with a yawning gap in expertise and ingenuity which will be hard to fill. He will be sorely missed by his fellow curators, but his legacy lives on in every nook and cranny of the museum.

Book Review : The Web before the Web – Ian Ritchie

Users of the World Wide Web are very familiar with the use of hypertext links that take them from one page or document to another and most probably assume that this facility was co-invented with the WWW itself. Not so, says Ian Ritchie in this fascinating and very personal history of his company OWL (Office Workstations Ltd). OWL was started in 1984 by a group of former ICL Dalkeith employees who took redundancy rather than move to ICL Kidsgrove. OWL developed a product called Guide, based on Professor Peter Brown’s work at the University of Kent on the on-line version of the PERQ workstation manual. OWL’s Guide was a very successful hypertext system that was sold to numerous companies with large product manuals that were being computerised. The road to success was pretty bumpy, however, and as readers we are treated to lots of anecdotes and insights as we are taken along it.

In 1990 Ian was approached by a young man who was working on an unfunded project at CERN. Tim Berners-Lee was hoping OWL could create a browser for his WorldWideWeb [sic] but with no money to pay for it and with no obvious market (the Internet barely existed) Ian had to decline. Hindsight is a wonderful thing.

I can heartily recommend this book. The text itself is a bit bumpy in parts but I only found one factual error, in the Glossary, where MU5 is described as being micro-coded – it wasn’t, as Ian is now aware.

Letter to the Editor

I noticed in Resurrection 98 a reference to the Arcturus A18D mini-computer.

I worked on an Arcturus A18D which was purchased in 1973 and was located in the Atkinson Morley Hospital, Wimbledon. The hospital specialised in neurosurgery and was the place where the world’s first CT Scanner (the EMI Scanner) was tested.

The Arcturus minicomputer was purchased to carry out further manipulation of the images produced by the CT scanner. It had 16K of core memory, two disc drives (2MB each, I believe), a magnetic tape drive and paper tape. As the unit was intended to display medical images, a 6400 word block of core memory was assigned to drive an image on a monitor. Here is a photograph of the setup.

On the left you can see two screens. The right hand screen is a VDU (ITT I think). The left hand screen is the monitor used for displaying the CT images. In front of it, there is a small box which was a joystick. I seem to remember that it worked by outputting different ASCII characters according to the direction. The box above the VDU is the electronics which converted the 6400 word memory block into the video input to the monitor. I think there were some brightness/contrast knobs. It was advanced image processing for its time! A piece of useless information: I think that the ‘18’ in A18D was the actual word width – 18 bits. It was a 16 bit word with 2 parity bits.

A year or two after the computer was purchased, Arcturus went bankrupt, but the same people continued to provide service in a new company, which I believe was called Leigh Systems Ltd.

I was employed as a research assistant, and I programmed the minicomputer. The manufacturer supplied very limited software, an assembler, linker, a line editor and a BASIC interpreter. There was no software for handling the discs, so my first job was writing a rudimentary disc operating system. All the software I wrote was in assembler. The system I wrote input the CT images from magnetic tape produced by the scanner, and allowed various image manipulation procedures, such as “stacking” the horizontal slice images, and producing vertical “slice” images.

50 Years ago .... From the Pages of Computer Weekly

Intel 8080 is ten times faster: Leading microprocessor manufacturer Intel has announced a new product, the 8080, which is compatible with the established 8008 but is about 10 times faster. A quantity price of $360 (£150) has been quoted in the US. The new microprocessor is a fully parallel eight-bit unit with an instruction cycle time of two microseconds, and a repertoire of 78 instructions. It is a 40-pin device with 14 control lines, an eight-line bidirectional data bus and a 16 line bus for addressing the memory and for input/output selection. Through this bus system the 8080 can access directly up to 64K bytes of memory, operate up to 256 input and 256 output channels and handle up to eight interrupt levels. (CW 391 2/5/1974 p7)

Plans for test runs of TV data transmission: A unified national system of data transmission using domestic television receivers has now been agreed between the BBC and the IBA and experimental transmissions are planned for the autumn. The system incorporates features of both the BBC’s Ceefax, announced in October 1972, and the IBA’s Oracle, which came out a few months later. The principle of both systems is to make use of suppressed lines in the video transmission to carry data which can be displayed on the television screen by means of a special adaptor. The sponsors of the experiment envisage that the service could eventually be used for broadcasting such items as stock market prices, sports results and weather forecasts, which could be frequently updated. Since it would be a low cost medium, it could also carry classified advertisements and entertainment guides. (CW 392 9/5/1974 p23) )

Argus aids Nottingham traffic flow: More details have been released of the dual Ferranti Argus computer system to be used for controlling traffic flow in Nottingham. Supplied as part of a £250,000 contract with Ferranti that includes traffic light controllers and Pelican crossing equipment, the two Argus 700E computers, to be located at a traffic control room in the city centre, will each have a 41K memory and 10 million words of backing store on disc. Peripheral equipment will include alphanumeric VDUs for use by control room staff and thermal printers for producing permanent records. The Argus 700Es will also be linked to an auxiliary control room at the Nottinghamshire Combined Constabulary communications centre at Egglestone and to the local fire service headquarters. (CW 392 9/5/1974 p23) )

MOS-memory minis launched: Continuity is the theme of new announcements this year, with the emphasis being placed on preserving the customer’s peace of mind by designing new systems to be completely compatible with the manufacturer’s existing product line. Announcing the first two of its new 21 MX series of computers, Hewlett-Packard made it clear that it does not intend to stray from this trend, stating categorically that the two machines, the 21-M/10 and the 21-M/20 are fully compatible with all existing Hewlett-Packard minicomputers. Seen for the first time anywhere at Hanover last week, and scheduled to make their UK debut at Eurocomp, the new minis will not be seen in the US until next month. The 21 MX series does not impact at all on the 3000 series, but will eventually supersede the 2100, which will nevertheless remain in production for at least 18 months. The new machines are aimed primarily at the OEM market, where Hewlett-Packard does more than half its business, and orders for delivery from August will at first be taken only from customers wanting five or more machines. End-user deliveries should start in January, 1975. The two new user-programmable 16-bit minicomputers have a newly-designed central processor, memory unit, power module and mainframe, the input-output channels being the only feature unchanged from the 2100. (CW 392 9/5/1974 p29) )

GEC equipment for Underground project: A £700,000 project, financed by the government and the Greater London Council, to improve services on the Northern and Victoria lines of London’s Underground railway system has been announced by London Transport. As the first stage in the project, a £300,000 order has been placed with GEC-General Signal Ltd for a dual 192K byte GEC 4080 installation, together with two 4. Megabyte cartridge disc units and associated equipment from the GEC March 4 series, for railway control purposes. Scheduled to be installed during 1975 at the Euston control centre for the two Underground lines, the control system will be developed over a five-year period. When fully operational, the system is expected to provide a regular service, particularly on the Northern Line, by making the best possible use of available staff and rolling stock. One of the 4080s will be programmed with data including train timetables and crew duty schedules, and will also take on the logging of the movements of up to 140 trains. VDU terminals will be installed at the Euston control centre to access the 4080 for information on train and crew positions at any time. In addition, depots and crew relief points will be fitted with online teleprinters. (CW 393 16/5/1974 p3) )

Com-Share’s DBMS aids navigation system evaluation: The latest series of efficiency evaluations to be performed on Decca Navigator’s Omnitrack air navigation system is to be aided by the use of the Composit 77 database management system, provided by Com-Share’s London bureau. Omnitrack is based on a Decca minicomputer carried on board the aircraft. This operates a moving map display, according to data from the aircraft’s internal instruments, thus enabling its position to be deduced without reference to signals from ground stations. The system can also drive the automatic pilot. Decca claims that the use of Omnitrack results in a saving of fuel and a simplification of air traffic control. The main purpose of the current assessment is to test the accuracy of these claims. To this end, a detailed comparison is to be made between air routes equipped with Omnitrack and those not using the system. The data obtained will be collated and analysed using Composit 77. Com-Share anticipates that the eventual size of the database will exceed two million characters. (CW 394 23/5/1974 p12) )

1903T/2903-based medical project: Following local government reorganisations in April, the new Mersey Regional Health Authority, in Liverpool, is to undertake an ambitious medical project based on an ICL 1903T front-ended by a 2903. The installation will eventually replace an existing 1902 system in Liverpool and have a wide range of medical and administrative applications. Scheduled to be installed in a new computer centre next January and to go live in April, the 1903T, with 96K core memory, will maintain a service for a large region around Liverpool, from Southport to Macclesfield and Crewe, including the Wirral area. Major applications to be supported include payroll for about 50,000 health employees; accounting; hospital activity analysis; and hospital information systems for three large hospitals using offline STC terminals. The 2903, which is expected to be delivered at the end of next April, will be used in an experiment, supported by the Department of Health and Social Security, and run by the Liverpool Area Health Authority (Teaching). In all, 16 VDUs will be located in an adjacent, new medical teaching centre, scheduled for completion in early 1976. The 2903 will be used partly to test out the viability of running small machines in hospitals, which are supported by mainframes such as the 1903T, for file handling work. Other uses for the 1903T include radiotherapy dosage planning; medical record keeping; and the support of an experiment, again under the auspices of the DHSS but run by the MRHA, into the use of computers in pathology areas. (CW 394 23/5/1974 p13) )

PSI aids colour TV: Special software written for a Digital Equipment PDP-8 by Programming and Systems Implementation Ltd has produced improvements in inspection routines for high precision shadow masks used in the manufacture of colour TV tubes at Thorn Electric’s Skelmersdale plant. The machine used for inspecting the masks is a Ferranti Cordax unit which incorporates a specially designed probe to move very precisely in x, y and z axes under the control of the PDP-8. This measures the minute and critical distances between the 440,000 holes in the mask. The location of each hole determines the exact position where the three electron beams, corresponding to the primary colours red, blue and green fall on the television screen to produce the colour picture. PSI says the software it has written for the PDP-8 has increased the number of holes for which co-ordinates can be recorded by the computer from 98 to 1,000 and sequential operation has reduced the computer time for testing each mask by 50 per cent. (CW 395 30/5/1974 p11) )

UK police databank goes live: Two years after the installation of a twin Burroughs 6700 system, the first phase of the Police National Computer project has been completed. This is regarded as a milestone by both Burroughs and the Hoskyns Group, which has been responsible for software development. The computers, which are situated at Hendon, in North London, form the heart of a nationwide network giving police forces throughout the country access to 5,000 million characters of information via 800 VDUs and associated printers, 300 of which have been installed so far. It is believed that this central file constitutes one of the largest single on-line databases in civil government use, and demonstrations on the Home Office stand at this week’s Communications 74 exhibition and conference in Brighton will show how the network will help the policeman on the beat. (CW 396 6/6/1974 p1) )

Ultrasonic aid in fight against cancer: An ultrasonic scanning technique for the early diagnosis and monitoring of treatment of cancer has been developed by scientists at the Institute of Cancer Research working with doctors at the Royal Marsden Hospital, Sutton, Surrey. At present the sound waves from the ultrasonic probe are converted into an echo picture which is displayed on a monitor screen and photographed with a Polaroid camera for further study. But only a fraction of the information contained in the sound stream can be accommodated on the screen, and the team plans to use a Digital Equipment PDP-8/e mini to record and process the data more fully. Analysis of such parameters as the scatter and attenuation of the received sound is ultimately expected to enable doctors to make a very precise determination of the nature of the various areas of tissue revealed by the ultrasonic probe, which produces a series of representations of “slices” through the organ under study. Soundwaves, being a form of energy, can be harmful in large concentrations, but the technique used at the hospital requires less than a thousandth of the minimum dose which could be harmful. In addition, unlike X-rays, harmless doses of sound waves have no cumulative effect, so that the technique can be used regularly and frequently on a patient undergoing treatment with no danger at all. (CW 396 6/6/1974 p40) )

NPL, Iria networks in Anglo-French link: The first test of “host-to-host” protocols between the National Physical Laboratory’s computer network and the French Cyclades network are expected to be made next week. A CTL Modular One on the NPL network is linked via Codex modems and a 9,600 baud line to one of the five CII Mitra 15s which make up Cigale, the Cyclades network switching system. Communication with the Mitra 15 mini, which is at Iria, the Institut de Recherche d’lnformatique et d’Automatique has been at the lowest level using the IBM BSC line control protocol, which has been adopted for Cyclades because of the IBM machine on the network, and on top of that packet formats have been transmitted. Some of these dummy packets, or “bubbles” as the French call them, have been dropped into the Mitra 15, while others have been sent with NPL as the address and these have been returned by Cygale. (CW 397 13/6/1974 p1) )

Laser-based system from InterScan: Still hoping for an upswing in the sluggish UK market for OCR equipment, InterScan plans to introduce a completely new laser-based system here next month. Known as the InterScan Laser One, it will compete directly with the IBM 3866 in the low-to-medium price bracket. At the same time, InterScan’s German subsidiary has sold a Scan Data 2250 reader in Switzerland which the c ompany believes to be the first anywhere in the world to read all 26 alphabetic characters in handprinted form in a full-scale live data capture operation. The Laser One is manufactured in the US by Optical Business Machines, of Melbourne, Florida, and about 25 systems have been installed and another 20 are on order. The machine reads characters using a laser beam scanner at a speed of about 350 ch/sec. Models can be supplied to read either the OCR A or OCR B fonts, plus handprinted numeric characters. Sizes of forms handled can vary from A4 down to 5 by 3 inches, single line documents or stubs. (CW 397 13/6/1974 p56) )

Microprocessor bid by Ferranti: In a bid to stake a claim in the world microprocessor market which, it is estimated, will be worth £500 million by 1980, Ferranti is developing its own “computer on a chip”. The F100L, as the new unit is known, is due to become available in 12 to 18 months’ time at a price of about £150 for OEM quantities. Manufacture of the 16-bit microprocessor will utilise the collector diffusion isolation LSI technique developed by Ferranti for fabrication of integrated circuits. Basically, collector diffusion isolation is a simple, high-yield bipolar process which requires only five masking operations, offering the function densities of MOS and related technologies. The logical design of the F100L was carried out by the digital systems division of Ferranti, at Bracknell, Berks, and trials and manufacture are being handled by the electronic components division, at Gem Mill, Lancashire. A spokesman for the company, at Gem Mill, told Computer Weekly that a trial chip had been nearly completed, and a TTL gate-for-gate equivalent had been made, for software development and rigorous testing purposes. The F100L is a single address, single accumulator fixed word length central processor with a 16-bit word length with twos complement fixed point arithmetic. In addition, direct addressing and indirect addressing, with or without automatic indexing, are provided. (CW 398 20/6/1974 p1) )

London link to Arpanet enables otherwise impossible UK research projects to get off the ground: The London link to Arpanet is now in its eighth month of full operation, and at a Users’ Meeting in mid-March, no fewer than 50 representatives from 16 universities and 13 research establishments attended. Already several research projects have been started in the UK which would not have been possible without access to the network. So far 25 applications have been approved by the UK Governing Committee, which consists of Professor Peter Kirstein of University College London, Donald Davies of the National Physical Laboratory, and Roy Bright of the Post Office. Funded by the Advanced Research Projects Agency, Arpanet currently links over 60 host computers all over the United States, including Hawaii, and these range from the giant Illiac IV, now at Ames Air Force Base in California, down to Nova minis. In the UK the IBM 360/195 at the Rutherford High Energy Laboratory is also now a host on the network. UK access to Arpanet is via a TIP, Terminal Interface Processor, loaned by ARPA and installed at University College, London. The London TIP is linked via a line provided by the Post Office to another TIP at the Norwegian Seismic Array Centre, Norsar, and thence by transatlantic cable to a TIP in Washington. An experiment planned to begin in the summer will offer an alternative route, via satellite. SIMPs, Satellite Interface Message Processors, will be installed at Goonhilly and at the US ground station, and connected by a 56kbaud digital channel operated in broadcast mode. Full details have yet to be worked out, but the British Post Office insists that the line to Norway must be maintained for the duration of the experiment. (CW 399 27/6/1974p24) )

The OCR wand casts its spell on the PoS market: The low-cost OCR wand, based on LSI technology, which is being developed by Recognition Equipment, has a wide range of applications with terminals, point-of-sale units, and is also expected to be of interest to banks. The wand is based on a 2.5 mm high solid state LSI chip in its nose which incorporates a 16 by 38 matrix of light sensitive elements. The character being read is illuminated by two adjacent light bulbs and the reflected light stimulates a pattern of electrical impulses in the LSI matrix. These impulses are accepted by electronics in the wand handle which streams each of the impulses into one of 16 voltage levels, corresponding to one of 16 shades of grey. The wand is linked via a flexible cable to additional electronics, which can be housed in the parent terminal, and which carry out such functions as character recognition using the feature analysis technique, and the assignment of an output code acceptable by the terminal’s logic. (CW 400 4/7/1974 p25) )

Laboratory on the right the lines: A mobile laboratory, which includes Hewlett-Packard equipment, is currently being used by British Rail in experiments it is conducting on Inter-City express trains. Known as Test Coach Mercury, the laboratory is being used between London and Birmingham. If successful, the experiments could form the basis of a system which would display information to drivers about the running of their trains. Initially, 30 Plessey transponders have been fitted on the northbound Euston-Birmingham track, and by the end of the year about 150 units will have been installed to cover operation in both directions. As Mercury passes over the transponders, signals are transmitted from an underfloor aerial. On receipt of a signal, each transponder generates a unique code in the form of binary coded decimal numbers, which can be used to represent messages, such as a position code or a speed restriction. The returning signal from the transponder is fed in to processing equipment and the resulting code passed to an H-P 2116B minicomputer. This then checks the data received against a known sequence held in store, and the results recorded on paper tape for later analysis. (CW 401 11/7/1974 p10) )

Micro16V base system to help fight diseases: A nine-year project undertaken by Dr Ken Bagshaw, head of the Oncology Department at Charing Cross Hospital, London, has resulted in a minicomputer-based system, the ramifications of which should have a profound effect on the treatment of malignant diseases. Developed jointly with minicomputer manufacturer Digico of Stevenage, Herts, the on-line automated radioimmunoassay analytical system, known as Darias, is based on a Micro 16V processor with 16K store. Radioimmunoassay is a special technique for the precise measurement of specific substances such as hormones, tumour antigens and drugs which are found in minute concentrations in biological fluids such as blood and urine. The technique involves adding radioactive labelled antibodies to samples of such fluids, successive dilutions of the solutions and measurement using a radioactivity sensor. Both the input and dilution unit and the radioactivity sensor device are under the direct control of the Micro 16V. Once samples have been diluted, they are filtered onto a special flexible plastic tape, the assay tape. This is then placed in the sensor device which executes a counting function in conjunction with a paper tape produced at the sample insertion of the assay. (CW 403 25/7/1974 p2) )

Saab-Scania units for Ariane rocket: Under an ESRO contract worth between £600,000 and £1million, the Aerospace Division of Saab-Scania is to supply nine special purpose computers for the Ariane rocket, a 200-ton three-stage launch vehicle, which is the biggest ESRO project so far. Ariane has been developed by a consortium headed by CNES, the French national space agency, with the aim of producing a European launcher to put up geostationary telecommunications and research satellites in the 1980s. The computers are Saab OBC/AR modular space machines designed for ultra-high reliability. They will be installed on board the rockets to control manoeuvres in flight such as attitude, direction, inclination and height. Designed for high reliability, the OBC/AR was developed under an earlier ESRO contract placed in 1972. It is a compact and robust unit weighing 10 lb and will handle real time and multiprocessing functions. The modules include bus controller, processor, direct memory access and interface buffer. (CW 404 1/8/1974 p1) )

Air traffic control data handling improved: An attractive alternative to the use of large-scale central computers for data handling and display systems in air traffic control, air defence and other communication and control systems, has been introduced by Marconi Radar Systems, a GEC subsidiary. Called Locus 16, the system makes extensive use of distributed processing in a way which makes even a conventional minicomputer seem inflexible. As a result, designers are able to build into each control point of a network all the data processing and storage required to meet the needs at that particular point, and have the freedom to design each control point separately on a one-off basis where requirements vary from one control point to another. There are several benefits to be derived from such an approach. The redundancy which needs to be built into such a system can be limited to one or two Locus 16 units, whereas in a centrally computer-controlled network, a mainframe on stand-by is normally needed. If a central computer is required, it can be much smaller than would otherwise be necessary, and would not cause the network to go down if it were to be put out of action. Any number of Locus 16 units can be interconnected using two-wire serial links. In a typical air traffic control system Locus 16 units would be installed at the radar sites to receive input data, and at each operator position. Each unit would be configured according to the local application, by plugging in appropriate circuit board modules – processors, display handlers, communication handlers and storage. (CW 405 8/8/1974 p3) )

Algorithms improve EMI-Scanner system: Improved algorithms are the key to a four-fold improvement in picture resolution on the EMI-Scanner, the computer-aided brain X-ray system built by EMI around the Data General Nova 820 mini. Using the new algorithms, which were developed at EMI’s Central Research Laboratories, a considerable improvement in cranial diagnosis has been achieved, says EMI, particularly in the field of orbital examinations which cover investigations of the eye, the optic nerve and the tissue behind the eye. This is because using conventional X-ray techniques it is necessary to inject radio-opaque substances into the optic region to improve definition, subjecting the patient to considerable risk and demanding exceptional skill in the control of the X-ray dose. With the new package, optic nerve and eye-structure features are clearly visible. (CW 405 8/8/1974 p9) )

Radar simulator based on PDP11: A radar simulator system, the SY2084, based on Digital Equipment PDP-11 minicomputers has been announced by Solartron-Schlumberger, of Famborough, Hants. Depending or customer requirements, the Solartron equipment is built around any unit of the PDP-11 range, with total system costs ranging from about £50,000 to over £1 million. Current applications for an SY2084 are in air traffic control and marine navigation simulation but, the company claims, potential applications are passenger check-in and baggage handling work. The system incorporates a Solartron-designed interactive display keyboard terminal which allows an instructor to communicate directly with the PDP-11. Software held in the PDP-11 was written by Solartron. In addition to the terminal unit and PDP-11, the system also contains two other new products, a digital video map and a plan position indicator display which can be used for both simulation and operational radar applications. (CW 407 22/8/1974 p48) )

Low-cost two-chip microprocessor: A two-chip microprocessor which, the manufacturers claim, promises to be the most cost effective such system available, is to be launched next year by Fairchild Camera and Instrument Corp. Called the F8, and using Fairchild’s isoplanar technology, its minimum configuration comprises a central processor chip and a read-only memory chip, both of which include I/O ports and clock generators to reduce the need for external components. The CPU chip has 64 bytes of random-access memory and on-chip interface circuitry, which, says Fairchild, permits several functional variations to be made by software changes only, instead of the hardware changes that have to be made in other microprocessors. Likely areas of application for the F8 include point-of-sale terminals, intelligent peripherals, process controllers, major appliance controls, test and measurement equipment and communications systems. (CW 408 29/8/1974 p7) )

Forthcoming Events

Members and others are welcome to attend CCS Seminars but these are also available via Zoom.

London Seminar Programme

London meetings take place at the BCS — 25 Copthall Avenue Moorgate EC2R 7BP starting at 14:30. The venue is near the corner of Copthall Avenue and London Wall, a five minute walk from Moorgate Station and 10 from Bank.

You should use the BCS event booking service to reserve an in-person place at CCS London. Go to www.computerconservationsociety.org/lecture.htm . The service must be used for remote attendance.

For queries about meetings please contact the CCS meetings secretary, Roger Johnson at

North West Group contact details

Computer Conservation Society

Aims and Objectives

The Computer Conservation Society (CCS) is a co-operative venture between BCS, The Chartered Institute for IT; the Science Museum in London; the Museum of Science and Industry (MSI) in Manchester; The National Museum of Computing in Bletchley and the Bletchley Park Trust.

The CCS was constituted in September 1989 as a Specialist Group of the British Computer Society (BCS). It thus is covered by the Royal Charter and charitable status of BCS.

The aims of the Computer Conservation Society (“Society”) are:

• To promote the conservation, restoration and reconstruction of historic computing systems and to identify existing computing systems which may need to be archived in the future;
• To develop awareness of the importance of historic computing systems;
• To develop expertise in the conservation, restoration and reconstruction of historic computing systems;
• To represent the interests of the Society with other bodies;
• To promote the study of historic computing systems, their use and the history of the computer industry;
• To publish information of relevance to these objectives for the information of Society members and the wider public.

Membership is open to anyone interested in computer conservation and the history of computing.

The CCS is funded and supported by a grant from BCS and donations. Some charges may be made for publications and attendance at seminars and conferences.

There are a number of active Projects on specific computer restorations and early computer technologies and software. Younger people are especially encouraged to take part in order to achieve skills transfer.