RAPC Electronic Accounting Development Unit

Worthy Down site was handed over to the Royal Army Pay Corps in 1960 and became home to the Electronic Accounting Development Unit (EADU) who housed their computer centre with its IBM 705 system in the purpose built three-storey Slater House. The computer was used to process the 1961 UK census electronically for the first time.

Donal Weaver has provided us with his recollections of this ground-breaking unit:

EADU – In the Beginning

By

Donal G Weaver

It’s always best to start at the beginning and life at EADU started for me while serving with RPO Ashton-under-Lyne in 1957 (or was it 1958?).I was told to report to IBM in London.No reason given, no explanation, just a Rail Warrant and instructions to be there in two days time.IBM – who or what were they?

At that time I had accumulated some four years with RAPC having transferred over under the Pay Sgt scheme and spent most of that service as S/Sgt in the Pay Office at Eaton Hall Officer Cadet School, Chester.IBM, never heard of them!

On arriving in Wigmore Street and having found the hallowed offices of IBM, still wondering why and for what reason, I was greeted by other RAPC personnel with varying ranks from Captain to Sgt who all, without exception, had no more idea than I as to why we were gathered there.By 10 o’clock that morning we had the answer.We were to be interviewed and given an aptitude test, to decide who was suitable for computer training.Computer? What was that? A machine system that IBM had and were about to sell to the RAPC!

One hour to solve, individually, a variety of logical problems – was it seven or eight in all – then a late lunch, goodbye, and return to your RPO/Unit immediately. A week or ten days later, a posting to Devizes - apparently having passed the test –and where others who also had passed now gathered.In total I believe there were about ten of us.Three Captains, Peter Brenchley, Peter Louth, and Tony Matthews-Williams (ex RAF) two EO’s from the Civil Service, and five senior NCO’s.Hirst,Driscoll, Muldoon, Wilson and myself.A new Unit, an elite unit (or so we thought at the time) but with no home of our own to go to.

Someone, somewhere, was guiding us.Deciding who would do what, where, and when.IBM gave us courses on “absolute machine language, flow charts, etc.Some went back to Wigmore street for further training, some for “hands on” experience at various industrial/commercial establishments, and I, together with a certain Sgt Jock Driscoll, found myself at Bristol Aircraft and being fascinated by a vast contraption called a “Rolling Total Tabulating Machine”. A machine which seemed to have a never ending appetite for punched cards and which displayed numeric characters six inches high across a screen (actually a board with windows} some ten feet wide. With pulleys and weights at either side to produce the “rolling total”” affect this strange machine dominated the room - Jock and I had never seen anything like it. But then we had led a very sheltered life in the RAPC and never heard of Hollerith. We commuted daily Devizes to Bristol for about six weeks, learning nothing about Computers per-se but we did acquire considerable knowledge about the production and handling of Punch Cards.

After the Bristol experience we settled back into Devizes where “the Unit” now proudly had working accommodation.This comprised of three huts – one for Programmer/Systems personnel, one for Input Design and Production procedures and the third for Courses (mostly run by IBM). We now had more personnel. The original ten had multiplied to about twenty - two or three from the Civil Service and, significantly, WOI Pat Lilley had also joined our ranks, together with a few more Sgts.

At this stage most of us had still never seen a computer but were receiving basic courses on Machine Language for the IBM 705 together with ALGOL 58, COBOL, and a variety of other languages now long forgotten. Our mysterious guiding force now fine tuned the unit personnel into their designated tasks. Some would become Programmers, some Systems Analysts, others would make up the Input and Production Control Team. For a time rank became irrelevant to the tasks at hand.Sgt Brian Hirst was acknowledged to be the best programmer and headed that team. My apparent systems skills placed me as the Head of Input and Production Development (which included some programming). But there was dissension in the camp. Certain officers considered rank must take precedent over ability, and some senior NCO’s felt that this “computer lark” was not for them. So changes came about.

A few of the original team were posted back to more normal duties and others, having passed the same original aptitude test, were posted in as replacements. Jock Driscoll departed and in his place on the Input & Production Team came Jim Muldoon. Jim, like Jock, was a Glaswegian of Scots/Irish descent and had served in Cyprus. (I mention this as it becomes relevant later when we come on to “Life after EADU)”. Jim was a programmer. The three of us had a common background in that we had each started our military careers as National Service in Infantry regiments - Jim with the Inniskillings, Jock with the HLI, and myself with the Black Watch. Jim and I became experts with an eight Iron, playing at lunch-times on an improvised, one-hole course within “the playing fields of Devizes”.

During this period other things were happening in the background with a full overhaul of the Pay and Allowances Policy for Other Ranks where, when looking at rules and procedures, several anomalies were found. The classic one being linked to “Payments for illegitimate children”. This could be made by a soldier to the mother of the child but there was confusion as to whether a mother, serving in the WRAC, could make a similar allotment to the father of her child. Matters of this nature had to be clarified prior to Systems and Programming being finalised.

Some programs were already ready for testing – but where to test them – for IBM had no 705 computer in the UK at that time. The answer - at an American Military Base near Frankfurt. Programmers, in particular Brian Hirst and his team, took the Harwich - Hook of Holland route and departed for a week or two at a time taking with them their own work for testing and also programs for test from the other teams involved. On their return re-writes and corrections were made and re-submitted for the next trip to Frankfurt. This routine went on for several months during 1959/1960 and coincided with our move from Devizes to Worthy Down.

Now we began to feel a difference. A purpose built building of our own, a guardhouse and a guard at the gate, these things were what we recognised, felt at home with, and our number rose rapidly with conventional staff performing conventional duties.One exception to this was the appointment of an auditor – a S/Sgt keeping an eye on the costs - unheard of in a normal Unit.

Systems Analysis and Programming continued, testing went on, and then one day in 1960 we were introduced to the mysterious guiding force, the person who was heading up the whole project. Col. Slater, ex RTR Commander and now a senior RAPC officer. A “pep talk” on what was still required with a time frame, something we had never had until then - and immediately felt was unrealistic - until we understood that here was an Officer of Staff Rank who apparently always reached his targets, on schedule, and that schedule would be adhered too. After he had got that message across there was a brief “thank you” for what had been achieved so far and the announcement that the seven or eight remaining members from the original group that formed the nucleus of EADU had all been proposed, seconded, and awarded full membership of the British Computer Society, MBCS after our names! Then came, what at that time, appeared to be the bad news!

As punch card input feeds to the IBM 705 was a slow operation the decision had been taken to switch to magnetic tape. This decision coincided with the release of the IBM 1401 system. It may have been coincidental, but it did make sense, for magnetic tapes could be produced on the 1401 overnight and processed on the 705 the next morning. Considerable re-analysis of systems and re-writes of programmes resulted from this announcement, mainly because the original input source was still the punch card which now had to be encoded on to magnetic tape prior to being processed on the 705.

NB.Mohawk Data Systems, which in later years could do this as a direct input without the use of punch cards did not exist in 1960.

With the 705 and two 1401 magnetic tape computers installed, programme development became easier. No more trips to Frankfurt. Instead, the Production Control Team came into action. Programmes that were ready for testing on any given day were submitted by 1600 hrs - the Programmer(s) involved advised of the scheduled time their work would be run during that night and they had the choice of attending the run or being given result data the next morning. Many preferred to work overnight and, although not allowed into the main Computer Room, would watch impatiently, from the viewing area.

Being developed in tandem with the RAPC project, and using the same Computer system, was a Government project for the 1962 Census. In consequence, processing-testing time was at a premium in the first half of 1961 and several times created ill feeling between the two “sides” when testing over-ran allotted time causing later tests to be re-scheduled to the following day (or night).This 1962 Census project was staffed by the Civil Service – hence the two EO’s who started with us in 1958.59 - although they both came from RPO backgrounds.

Everything progressed well for a while with only the occasional “hiccup” – usually at night during testing run time – always due to “programming fault”, never (or seldom) would the operator on night-shift admit to any wrong doing. By June 1961 life was becoming more settled. Routines began to be established, programmes being signed off as completed and fit for service, and significantly some of us were gaining considerable experience - which had a market value in civilian life.This, I must confess, led me away from the military life I’d had since leaving Polytechnic education and into the unknown new world of civvy street…….so on August 1 ^st 1961 I became a civilian and changed from the Battle-dress to the Bowler and Brolly Brigade.

That original 705 console now takes pride of place in the museum at IBM Hursley Park near Winchester:

Original specifications of the IBM 705:

Operator's console: The operator's console is a separate unit, but is considered an integral part of the Central Processing Unit of the 705. The console is used to:

Control the machine manually.

Store information manually.

Determine the status of registers and counters.

Display the contents of memory and storage.

Typewriter: The typewriter on the Type 705 can be used to print directly from memory, one character at a time. The speed of typing is approximately 600 characters per minute. All other operations of the machine are held up during the typing operation.

The typewriter has a 12" carriage and is equipped with a pin feed platen. Maximum length of writing line is 8 7/8", overall form width 9 7/8", and hole-to-hole dimension is 9 3/8".

Any character not on the code chart prints as a question mark. A plus zero and a minus zero print as a plus sign and a hyphen respectively. Sensing the record mark automatically causes a carriage return and automatic spacing in accordance with the setting of the space control on the carriage.

Main power supply: The main power supply is furnished as a separate unit of the 705 and supplies the necessary regulated power for the operation of the Central Processing Unit.

Questions in the House of Commons!

Extracts from Hansard:

Computer Assisted Pay Accounting System

HC Deb 03 August 1962 vol 664 cc178-80W

 
178W § Mr. Wade asked the Secretary of State for War what has been the annual current and capital expenditure from the financial year 1953–54 to date on the Royal Army Pay Corps' electronic computer; what has been the value of lands and buildings appropriated; and what is the estimated annual cost of its operation.

179§ Mr. Profumo The total expenditure to bring the computer assisted pay accounting system into operation is estimated at £1,800,000, incurred since 1958.

180W The new building to house the computer at Worthy Down cost £225,900. The land it occupies is a very small part of an area taken over from the Admiralty. The annual running costs of the computer centre itself are estimated at approximately £350,000.

Royal Army Pay Corps (Electronic Computer Organisation)
HC Deb 03 April 1963 vol 675 cc441-2

441 §30. Mr. Bellenger asked the Secretary of State for War whether the electronic computer organisation set up by the Royal Army Pay Corps is being utilised by all three defence services.

§ Mr. Profumo No, Sir. The existing facilities at Worthy Down are sufficient only for the requirements of the Army.

442 § Mr. Wade Is the Secretary of State able to say what financial savings have been achieved by the use of the electronic computer organisation?

§ Mr. Profumo Not yet—it is too early—but it is being an extremely great success.

Worthy Down (Electronic Computer)
HC Deb 10 April 1963 vol 675 c159W

159W
§ 51. Mr. Bellenger asked the Secretary of State for War what was the capital cost of providing the electronic computer installation at Worthy Down; and what consequent saving in manpower there has been.

§ Mr. Profumo About £1,800,000. We have saved about 575 staff.