Dick Cromwell Leo memories – I went to LEO II at Elms House Hammersmith in February 1959. I left in April 1963 to start LEO III training. I started as an Assistant Engineer, then Shift Engineer and finally in July 1962 became Chief Engineer at the site. I moved on to LEO III/16 and took it to Kayser Bondor at Baldock Hertfordsire. Dick’s full account can be found at https://www.dropbox.com/s/fd5p9o24j8zh0rt/Dick%20Cromwell%27S%20LEO%20MEMORIES.docx?dl=0
https://www.dropbox.com/s/0s4nweb17m4zqwm/Anthony%20Robin%20Davies%20memoir.doc?dl=0
John Daines: Reminiscence
The Glyn Mills payroll paid people by Credit Transfer rather than by cheque**. It was at the start of each bank branch having a sort code as well as folk having account numbers. We produced a payslip and a credit transfer (Credit Advices on the pic). There was a tape that contained “the Bank File” and that started to be used for more payrolls – “the standard payroll” that became an early package, requiring only “minor” tailoring for each customer.
They paid military officers (all army, half RAF or half army, all RAF I think and the input data was referred to as ”casualties”, a term used in the pic. I seem to remember that the casualties were punched and verified twice and two tapes were input for comparison (Bob Stevenson may be able to confirm).
Glynn Mills had their offices at Osterley Park (wartime evacuation) and I remember that on one weekend I had to deliver the results to Osterley. It was when the Hammersmith flyover was being built so it involved a lot of in-and-out under the works on the A4.
I did some googling and discovered that they were also adopters of another technology used in an innovative way. See Wireless World Sept 52 page 379 Glynn Mills use of Television for documents
Ray Dawson, Reminiscences published in Bits and Bytes, Newsletter for ICL Pensioners. Autumn 2000, pages 3-4, and https://www.dropbox.com/s/pvn23zfn0p6ylwf/Ray%20Dawson%20Reminiscences.docx?dl=0
JD also remembers the Master Routine: I have listings of the master routine and it was written in Intercode.
Intercode itself was a level above machine code and, although a instruction looked to be an equivalent to a machine code instruction, it was often expanded by the translator into several machine code instructions.
However, Intercode instructions 100/0/0 to 131/1/3 were one for one equivalents of machine code instructions 0/0/0 to 31/1/3. That meant that the master routine programmers could program at the lowest level and use specialist low level instructions that weren’t in the Intercode set e.g. input output, interrupt handling, setting store protection tags .etc
Interestingly, Cleo allowed for routines to be written in Intercode and, by implication from the above, that Intercode might include machine code
John Daines November 2021
I’ve always assumed that P1, P2 and P3 are constituent programs of the L3 suite – Bakery Valuations.
From reading some of the files, these are very complicated programs and the machine was tiny.
The following shows the amount of time involved.
From looking at Lenaerts Notebook No 8 page 39 written on Monday, Dec 3rd 1951, he refers to the P programme’s successful run on Friday afternoon (November 30th) at 2:20. Note that Len then continues with his usual pre-occupation (faults!).
Note also that on the 28th, page 37, he notes the successful completion of P1, suggesting that perhaps P2 ran on the 29th.
Three days therefore looks a reasonable view.
John Daines LEO III Programming and Operating Utilities
An important utility managed program changes with the Version Control utility implemented in the Intercode Translator. A programme had a 5 digit version number that was incremented by 1 each time the program was amended. See Volume III, section15 at http://sw.ccs.bcs.org/leo/LeoIC6-17.htm#s15 15.1, 15.2 and 15.3 last paragraph “New Issue Number” Each line of code was numbered by the Translation when the program was first created and, if it had been amended, the program version number of the amendment was printed next to it. All this was built in. There was a program trials facility built into the Translator see volume III section 16 at http://sw.ccs.bcs.org/leo/LeoIC6-17.htm#s16
There were standard sort utilities 07003 (3 tape sort) and 07003 (4 tape sort). They were string sorts where the strings were created on two work tapes from the input tape in the first pass and subsequently merged, utilising the hardware merge instruction, until the final two strings were merged onto the output file.
Printing could either be directly to a line printer, which would restrict its use by other programs, or by sending each print line to a magnetic tape file with a header, that said what type of report it was, together with any paper movement controls. Thus a program could effectively output to any number of “printers”. The standard print program utility 06060 would subsequently read the magnetic tape, printing all lines of a report type on each pass. Lines of print were assembled by a table driven hardware instruction that selected data as required, changing formats and inserting £ signs as required. All this was cross checked by the Translator.
There were also utilities to manage program libraries, copy and compare magnetic tape files and print the contents of magnetic tapes and main store. See volume V.
More importantly, there is an extensive piece of work by Ken Kemp, who was in charge of Leo systems and programming at the English Electric-Leo Service Bureau in Bristol in the mid to late 1960’s. At http://sw.ccs.bcs.org/leo/Manuals.htm , the main index to the manuals, there is a pointer to Recollections of a Leo III user, which is at http://sw.ccs.bcs.org/leo/KenK.htm
Barbara Dickens, I was a LEO programmer in Intercode on the Leo 326 installed by the Dept of National Savings. I started in 1968 and was trained by the Post Office trainers before working for DNS at Lytham St Annes. I remember the days very well as I did mainly maintenance which meant finding old code to delete to make room for any amendments. After a few years I moved onto ICL System4 machines programming in Usercode and COBOL before joining SMBP and working in COBOL and machine code on Univac 1100 series.I eventually moved onward and upward into management and consultancy but I still remember my Intercode days and the carol tunes made from holes in the printers format tapes.
George Drummond, GPO (BT) –Reminiscence Born in Fife Scotland in Feb 1947.Attended Kirkcaldy High School, then Strathclyde University. Initially applied to the Post Office and was interviewed in Edinburgh for an Executive Officer post ( as they were called in those days ). During the interview I was asked if there were any specific areas in which I would like to work, and I said computing. I was then asked to go down to London to sit a computing aptitude test, and from there I was offered a job on the Telephone Billing programming team in Docos House, Aldgate London. I was the last BT in-house instructor to teach Intercode and CLEO programming. When the St Albans training school no longer had instructors in house, it was passed to the TB3 project team to take over that role, as we were the last operational user of those languages. We even had a few non-BT students on the later courses as , by that time, no-one else taught those languages. As a result of BT’s decision to offer this externally I had to go through a six-week course at BT’s training college to learn how to become an instructor and to become an accredited courses lecturer before BT could ‘Sell’ this training commercially. I started work originally with the Post Office in 1968 working on the Leo 326 Telephone Billing Suite of programmes known as MFC ( Major File Change). These were a suite of programmes used when Telephone Exchange Ranges were being altered to support the STD dialling system. The suite was capable of carrying out complete, or partial, exchange re-numbering. I also worked as a senior programmer on the introduction of Post Codes, Decimalisation programme, and VAT introduction for Telephone Billing. From several years I took charge of the programmers who provided Software Support to all of the operational Leo326 telephone billing systems. When BT then moved to the 2900 systems, I relocated back to Edinburgh ( my home town) as a System Software Support manager based in Craiglockhart. I subsequently moved into an IT sales technical and system consultancy role for the final four years of my 39 year career with BT.I retired in 2007 now live on the South Coast in West Sussex.
https://www.dropbox.com/h?preview=George+Drummond+memoirs.doc
John Edwards Reminiscence after attending Neville Lyons Lecture I worked only briefly for Lyons doing Time and Motion Study at Cadby Hall. It was in fact an interim job as I resumed my degree work at London University. I was making the most of a grant to study after wartime service as a navigating officer in the Merchant Navy. I got to know most of the departments at Cadby Hall, including T&M on the swiss roll line that you mentioned. I also did some studies at the Coventry Street Corner House. This was to determine the average time spent in the Brasserie ; this was done by simply counting the numbers going in, and those coming out , over a period of time . I remember a somewhat hilarious moment in the ice cream factory when a stainless steel overhead pipe carrying the liquid ingredients became disconnected and poured a fair amount over the foreman.
The overall impression I had of operations at Cadby Hall was great efficiency. The main point of the T&M work was to get control , by knowing the man- hours required to produce the various items. I think that efficiency was the main driving force behind Leo , was it not ? I did not have any direct connection with Leo, I was told that there were some people from Birmingham University working on it. I have always tried to impress on people the remarkable achievement of the first commercial computer being developed by a food company, and by the lead that this should have given the UK in that science. I feel that Lyons should have persisted in being in the computer business and given government support. I know from personal experience how little support was given to new technology. In 1971 I had a small electronics company, in that year Intel produced the first microprocessor (the 4004) we used this to produce the first desktop computer in the same year. Unfortunately in spite of getting orders from UK and abroad, and being used in commerce , lack of adequate finance killed it off. https://www.dropbox.com/s/tjlxl5xtiyd8p0d/John%20Edwards%20Reminiscence.docx?dl=0
Stan Evans, Maintenance Engineer on LEO I and LEO II, joined LEO 1956
Date of Birth: Born 23/09/26
I never realised that my life was about to dramatically change when l went to Joe Lyons at Cadby Hall in 1956. for an interview for a job as maintenance engineer working on their first computer they had ever made. I was an experienced electronic technician but knew nothing about computer digital technology.
I was interviewed by the computer engineering manager Peter Mann. He asked me many questions but sadly I did not do very well with my answers. To my surprise and pleasure he offered me a job. It later emerged that all the other candidates had the same lack of knowledge that I had.
What was unique about this computer was that it was the first computer “worldwide” that was designed to do commercial type work. The other computers in use were basically scientific machines.
Lyons formed a company called LEO Computers so in fact I joined this company not Lyons.
There were two other chaps who joined me the same day. Les Rabitt and Stan Holwell. Now this is where I go for a wobble trying to remember names. However here goes.
The senior engineers already on site. Frank Walker, Jimmy Wheeler, George Manley and a much older chap Chas. These were the chaps who not only had to mend LEO I but had to teach us rookies.
Although there were detailed schematics of every specific unit in the machine there were no written notes on the logic. ie how the computer worked other than the hand written notes of our tutors . I pause a minute to reflect “You cannot get more basic than that.”
Our work shop was a 12 ft by 12 ft room located across the corridor from the Computer Room. In addition to us engineers there were three other people in this room.
To help fixing computer faults every unit type had a spare unit. I think they were called LCU xx If a fault was traced to specific unit and the fault cleared when replaced by the spare then Hey Ho back to the operators and the machine was up again. In the work shop the faulty unit was put into the Test Rack.
This Test Rack had simulated features to be able to test the unit and find the fault. There was a very clever engineer called Gibson. He maintained this test rack and also was involved in the thinking about the next step LEO II. Also in the workshop were two ladies.
One was a very experienced lady who did all the electric wiring in the units (modifications etc) The other lady was a thermionic valve tester. She had a check list of units that had been operational for a certain time period and checked all the valves and if below the Makers specification replaced them with a new valve.
The huge irony was that a development engineer chap called Lenearts established that new valves were more likely to fail than the comfortable settled in old ones. So this testing of valves did not really do any good. So this procedure was stopped. We were learning all the time.
There were two people we often worked with on the computer. They were David Caminer and Mary Blood. They were in fact the first Business Computer Systems and Programmers in the country and David led a small team of programmers who had written all the LEO I operational jobs since 1952 which was quite impressive.
Programmes were written in machine code format and stored in the main memory. This consisted of a number of special long tubes filled with mercury. They were stored under the false floor together with a number of smaller tubes, registers that contained only one word of 17 binary bits. Note. I am struggling here and hope I have got it right. I would think the main memory had 32 tanks and each tank 32 words.
I seem to recall a D30 digit. Perhaps that was Leo II
Basically a line of machine code had 4 areas. Each area had a specific significant meaning. If the programmer wanted this action to be an ADD he would put a number in binary in the first area because he knew that number told the computer to ADD. The other three areas were locations of the two numbers in the action and where the answer was to go. Clearly this format would not be the same for other instructions.
How about that a programme lecture in one paragraph. The point of this paragraph was to illustrate how basic the programming was. Cobol, Fortran etc was not even a dream in these days.
Mr Caminers team had written several programmes which tested the performance of the machine. When it failed we were given the reason. If it did not fail we used the margins tester. We had over a 1000 wires from the engineers control panel going to the most crucial electronic sections in the machine. These wires sent electronic signals that increased or decreased digital pulse size. Use of variation of threshold often made the performance test programme fail. Using eliminating switches It was a simple matter to find what unit was causing the programme to fail.
In certain difficult faults Mr Caminer was a great help in typing in a simple programme which clearly showed what action was going wrong, the engineers were then able to get moving with the oscilloscopes and find the fault.
I have enclosed a picture of a young engineer all 30 years old Stan Evans himself working at the LEO I engineers control desk. It was here that the engineers and Mr Caminer did our fault diagnostic work. Very often the engineers would do the same for the same reason. A small loop of a failing instruction made oscilloscope work easier.
Interesting to note that the Operators rarely use the engineers panel because at that time operations control was done using Punch Cards. These cards had been successfully used for many years for basic business tasks. They were 3 inches by 7 inches, had numbered columns 1 to 80 with rows numbered 0 to 9. Rectangular holes were punched out in each column.
To start a programme the operator would push a button called BOOT which placed into the computer Order Tank an instruction to read a card into the computer. This card contained the necessary instructions to read the complete pack of cards into the computer. These cards contained sequence of computer instructions ie the programme, and the scheduled commercial programmes was then ready to run. The operators area was dominated by three big heavy card readers, equally heavy printers and a paper tape reader.
The inevitable happened of course the work load had got very heavy so we had to go to three shifts. I hated the night shifts and I think in hindsight we would have preferred One week of nights and two weeks day/evening. However we three new chaps were very rapidly promoted to Shift Leaders because the senior chaps were moved to Minerva Road factory where work on the next generation computer LEO II was moving very rapidly. Oddly enough I cannot remember anything about our junior engineers at that time.
Doing nights did have one advantage and it’s a story I was very fond of telling. The Lyons management who were very highly regarded and so were provided with an “Out of this world” restaurant. Two lady chefs looked after the restaurant at night. I used to chat these ladies up. Most of us, the engineers and operators enjoyed a very basic level of living so a full roast and three veg was a real treat. I was often offered if it was available a Dover Sole. I have never seen such a big Sole since then. A quarter of an inch fish each side of the bone and the fish overlapping the plate. This is how the rich or talented lived. I missed that!!!! when I eventually moved from LEO I So back to the real story.
Very soon the first prototype LEO II was completed and installed in a different but near bye building to Cadby Hall. Both Les and Stan had moved to Minerva Road but in spite of my pleading I was still left in charge of LEO I. Something about good training in Management for me. Eventually LEO II / I was up and operational taking the commercial work off LEO I.
Transfer at last. No training course slot available so straight in as Shift Leader with new trained engineer Tony Talmadge
to help me. (no the name is not right ) I knew LEO I so well that the experience was invaluable when thrust onto LEO II and Tony was good. Hurrah.
The concept was almost the same but the electronic engineering had taken a massive jump forward. Not only was the machine considerably faster it now had magnetic tape which meant good bye to loading programmes and storing data on clumsy punch cards. Delay mercury tubes were still used for the memories. So given time and not too many nasty faults I soon settled in to my new job.
However even then it was “Stan they are in trouble on LEO. I can you go and help them”!! I suppose I took it as a compliment.
I was on LEO II / I for about 18 months when I was asked if I would be willing to move to East London and install a computer LEO.II/9. Consult with family and Yes. I was also able to help commission the machine at Minerva Road with a chap called Peter Ross before taking it out and installing it at Ilford (the photo people). In the following years we became good friends but that is a Honeywell story.
What was unique about LEO II/9 was that the main memory .was now stored in a new device called a Core Memory instead of bulky mercury tubes.
I was now living quite close to Ilford which made a marvellous change not having to travel a long distance to get to work. I soon settled down and with a good crew I had very low down time.
I should have realised with my past LEO career it would not last for long. My friend Stan Holwell had been made Area Manager and I was asked to take up the appointment of Area Supervisor to support him and in particular address myself to problems at LEO II / 5 the in-house LEO machine located in Hartree House London.
There was a piece of new peripheral equipment being installed called Magnetic Drum but with good support from Minerva Road that ceased to be major problem and it soon became fully operation.
This was a Random Access Device (RAD) that enabled a programme to zoom in very quickly for data instead of a long wait going along a magnetic tape. Disc Drives did not arrive on the LEO II’s
Leo II’s were all thermionic valve machines with the inevitable problem of long periods of down time There was still quite a few difficult faults on Leo II / 5 that required me motoring from Ilford to mend them(Always late evening)But it was always satisfying to fix the problem and get back home again.
Reading back my report its interesting to note that the our memories are very selective. I read the Leo Computer Society members list and see that there were many chaps who I worked with and many others who sadly may have gone now who I do not remember.
One thing I do know Leo produced some of the most highly competent, dedicated engineers and regarded as the best in the country.
My career path with LEO finished quite abruptly for Honeywell Controls advertised for engineers to start up a Computer Division. These computers were the equivalent of the LEO III’s No valves fully transistorised. My golly what a nice change.
So after five years very exciting time with LEO I left and installed the first computer Honeywell had sold Bournemouth Town Hall. But that is another story.
To finish my Saga there was a humorous story about LEO II/9.
Many years later all the LEO II were being replaced with LEO.III’s Fully transistor techniques replacing the unreliable thermionic valves. The machines had little value to the Scrap Merchants except for all the mercury in the old memory tubes. One such dealer thought he had bought LEO II/9 for a song! When his gang arrived to remove the machine he found to his horror NO MERCURY it had a core memory. See also version with photos at https://docs.google.com/document/d/1EjtC3VTgqRgZ2HZGUwB4D9gbd2gcUNmm2aPHgMwmMME/edit and at https://www.dropbox.com/s/im70jnacidwe2jq/Stan%20Evans%20LEO%20Maintenance%20Engineer.docx?dl=0
Stan Evans Review of John Daines Zoom forum presentation on 14th July 2021 and reminiscence of TRT in action:
John Daines was brilliant . Not only the very professional way he gave the presentation but the many items of news I did not know. In particular those related to TRT in the war. Also Lyons running successfully a war effort factory.
One memory of TRT was in the middle of a nasty engineering fault on Leo I he came down and quite rightly wanted to know our progress. We had been down 3 hours by then.
He suggested you go into the work shop , have a cup of tea,switch off for 5 mins then recap the problem
Needless to say it worked. we were up again in 20 mins. A lesson I used many many times later on in my career
So I repeat John gave a brilliant presentation.
Of course feeding my ego he points an arrow at your truly at the Leo I engineers control desk in the section Maintenance