John McEneaney

This testimony begins in the ‘Ireland’s first computers 1956-69’ archive. Click here to view.

In mid-1974 I left UAC to join telecommunications service provider Bell Canada where I worked, initially, in network analysis and design of its internal networks. I then moved to the company’s minicomputer ‘district’, where my boss was Arun Slekys, a Lithuanian-Canadian genius (in my opinion) whose doctorate was on ‘trinary computing‘, a concept which I came across later when working for Honeywell. Arun was incredibly clever, had an almost fanatical work ethic and was a huge influence on my career at Bell. His mentoring (even though he is almost exactly the same age!) was of great benefit to me. In about 1981 Arun joined Hughes Network Systems in the USA. He later joined the board of C-COM Satellite Systems and chaired the Global VSAT Forum.

On my second day at Bell, I was determining what courses I needed. The head of the training unit, Roger Belanger, asked me if I knew Brian Rothery – who had, in fact, given me my first formal programming training!. It turned out that Roger and Brian had worked together and had many (often hair-raising) rock climbing experiences. On my return to my own desk, one of my neighbours asked me if I knew Aidan McKenna; of course I did – he joined Aer Lingus after it took over Independent Computer Bureau Services in 1972.

At Bell I worked, mainly, on Data General Eclipse minicomputers, in Assembler and FORTRAN 5, and in multi-mini ‘rings’ using Data General’s multiple computer adapter. This was a high-speed, fixed time-division multiplexer (TDM) bus which could connect up to fifteen separate DG Nova or Eclipse minicomputers, each processor getting 1/16th of the time ‘bandwidth’.

Arun worked with key designers from Digital Equipment to design a programmable TDM bus for the PDP/11 series. This supported Bell’s customer-facing automated loop testing system. It was the first time I encountered ‘virtual processors’ – in 1975. I have always believed that that Arun’s ‘bus’ was the forerunner of the VAX-bus, which was the basis of the multi-VAX systems implemented at the Department of Social Welfare.

I also came across Fred Brooks’ book, ‘The Mythical Man-Month’, on software engineering and project management. This became my project management bible. Notable quotes from it include ‘Adding manpower to a late development project makes it later!’ and ‘It takes nine months for a woman to have a baby; you cannot make it go quicker by putting two women on it!’

Talking about babies, we had decided that we would wait until whatever children we had were of school-going age before we decided if we would stay in Canada. Things did not work out as we intended, so we decided, in 1977 when our elder daughter was just over one year old, to finally settle in Ireland.

I had met Tom McGovern, then joint managing director of System Dynamics, in 1973 and, while we were home on holiday, Tom offered me a job with System Dynamics. I joined them in October 1977.

System Dynamics had installed Data General Nova minicomputers at Eason, Wiggins Teape, Brooks Thomas and a few other customers. I spent many hours, days and weekends at construction materials supplier Brooks Thomas, sorting out its computer problems and becoming aware of their stark commercial cost.

I also worked with Cyril Harrington, the data processing manager at Brooks, to test and implement a technologically advanced payroll system. This was developed by Nick Spalding – the one-time IBM field engineer that I had known at Aer Lingus – and there was very little for me to do. Nick was far too good a programmer for me to be able to find any faults!

Brooks Thomas had been having problems with ‘docket entry’ and invoicing and I made the stupid volunteering remark that one was similar to a hardware problem I had encountered on a DG Eclipse computer at Bell. It involved a difficult-to-trace hardware memory error. Also, memory 16-bit addressing limitations of DG’s RDOS operating system restricted the number of terminals that could be operated. RDOS had two address spaces – foreground and background – that could be used as (virtually) separate machines. This enabled us to double the number of terminals. By adding another multiplexer to a second DG Nova, Brooks Thomas could add more terminals and run other applications in ‘background’, albeit with some operational restrictions.

The only place where there was a Data General configuration sufficient to transfer Brooks’s data to the new configuration was at DG’s office in Birmingham. I was sent there just over a week before Christmas with 15 or more 10-megabyte cartridge drives and a couple of reels of tape, together with (incorrect!) customs documentation. I was the last person, including the guy who was locking up, to leave DG’s office on the day before Christmas Eve. It was still relatively soon after the Birmingham pub bombings, so I had to keep the 15 disk packs and boxes with me from shortly after noon until the flight left at, I think, about 7pm. Because they were on a ‘truck’ of the sort that delivery people used, I could not manoeuvre them to the next floor where the restaurant was located, so I had nothing to eat or drink until I arrived home shortly after midnight, having acquired the necessary documentation to have signed to release the disks, which had been impounded by Customs! I spent St. Stephen’s Day retrieving the disks and working on them at Brooks Thomas, to get the system running before the end of the Christmas break.

Early in January Tom McGovern met Cyril Harrington and remarked ‘Well, at least that problem has been solved.’ He made no reference to my efforts or commitment and did not even put his hand on my shoulder to say ‘well done’! At that exact moment, I decided I was leaving System Dynamics and in a matter of months I had joined Honeywell.

John during his Honeywell years.
(Photo source: John McEneaney. Photographer unknown.)

Honeywell Information Systems Ireland Limited was treated as a branch of the UK ‘northern region’. Locally, there was a (sort of!) senior management team, consisting of branch manager Kevin Treanor, systems manager Mervyn Kelly, who recruited me and later became sales manager, engineering manager Brian Howlett and Jack Scott, account manager for Honeywell’s largest customer, the Revenue Commissioners. There were about 50 or 60 staff, grouped by sales, technical (systems/programming) support and field engineering.

The sales side included Tom Ryan, who was the account manager at Norwich Union and VHI and later became the general manager of NEC Ireland, Bill Noble and Tony Flynn. On the systems side, Mike Keating was probably the most senior. He was the project manager on the implementation at builders providers Dockrells in Ballymount – a project that I joined at a fairly late stage. I think that Honeywell used it as training for me. Larry Duffy was a very senior and experienced project manager, who supported customers like VHI and Ballyclough Co-op. He and his wife started a ‘horse improvement’ farm in Shillelagh, Co. Wicklow. Larry left Honeywell, but continued to develop software, using the farm as his base.

The other technical people included Steve Craig (a member of Mensa with a huge IQ, who provided support to Revenue Commissioners but died in his 30s), Ken Jenkins, Frank Cloonan (who later joined DEC) and Brendan Johnson (who became the data processing manager at J&E Davy). Based in the Belfast sub-office were Malcolm Kennedy and Brian Baird, who subsequently became managing director of ICL Ireland.

As a technical adviser I worked mainly on the conversion of older systems from other manufacturers’ systems, such as Singer System Ten, Telex System III and IBM System/3. Those used a variety of languages, including RPG III, COBOL and Assembly languages. The discipline and methodologies I learned at Aer Lingus were very useful in doing system testing and in making sure that the converted systems gave the correct results – which were sometimes different from the results produced by the originals!

Almost all of these conversions were to Honeywell’s Level 62 small mainframe, which was renamed DPS 4 in 1979. Like all of the company’s computers it ran a version of the General Comprehensive Operating Supervisor (GCOS) or, colloquially, ‘God’s Chosen Operating System’! The Level 62 used, predominantly, COBOL, although we needed, occasionally, to use Assembly language called-routines.

In 1979, and for much of 1980, I ‘commuted’ to Shannon, working on a project at the EI Company, converting, primarily, production material management systems from a General Electric GE 110 – EI was itself a GE company – with its own (very old) operating system. The GE 110 used a sort-of RPG language called ‘Tab’ that we converted to COBOL, running on a Honeywell Level 64 and using the GCOS64 operating system. While classified as a small-to-medium mainframe, I considered the Level 64 (and its successor, the DPS7) the best minicomputer of the time, as the hardware and software seemed to have been designed as one!

On this project I came across, for the first and only time in a commercial system, the concepts of ‘positive zero’ and ‘negative zero’ which depended on the ‘direction’ of the calculation. COBOL recognised only zero or not-zero. We needed to replace a single ‘compare and branch’ combination with, initially, about ten lines of COBOL and (later and not by me) with a called routine written in Assembly language.

I took over support at bicycle maker Irish Raleigh when it was finalising the conversion of applications from a System Ten and beginning to develop a new online sales entry system. When the company’s data processing manager departed, Honeywell seconded me to fill this role until Raleigh could recruit a replacement. We devised a means of data entry using key-to-diskette and ran the production system in that way for quite some months.

Another project that I worked on was the development of a mortgage administration system at Norwich Irish Building Society. The detailed requirements changed during development – or more likely, were not properly expressed by them nor elicited by me during the analysis phase, so there was a very significant overrun, which was covered equally by Norwich Irish and by Honeywell.

Health insurance provider VHI operated a Level 64 computer. Its online system users were very happy with the system response, but batch operations – even initialising a tape – took, literally, hours. Honeywell had performance analysis software which inserted ‘probes’ into the operating system and into the interfaces between the operating system and the applications. These tools indicated to me that VHI’s IT staff had fine-tuned the system for maximum online throughput, at the expense of the less-time-critical batch functions.

I left Honeywell in late 1985 and joined Arthur Young as a senior consultant. We did a number of information engineering projects in the public sector and a number of consulting projects in the EU, working with Arthur Young (subsequently Ernst & Young) offices in London and Brussels. I worked with several very clever people from the UK and from Utrecht in the Netherlands. Information engineering was – to me, at least – totally new and it needed a PC with the highest spec available to run it. I also spent about 18 months working at Keating’s Bakery in Kanturk and helped them install a new bakery sales and production system, based on a PC network. I was especially pleased to have mentored Keating’s general manager who became a very effective IT projects manager!

In 1993 I joined the Irish League of Credit Unions (ILCU) as head of IT. I had been a credit union member since my time at Aer Lingus and had later been involved with starting one in Celbridge. I regarded this as my second job! In 1996 ILCU undertook a joint project with Bank of Ireland and installed ATMs in a number of credit unions. This might seem like a relatively trivial achievement, but it gave huge benefits to those credit unions and their members.

I returned to what had become Ernst & Young as programme manager on the Courts Service computerisation project. The company sold its consultancy division to Cap Gemini in 2000, but six years later Cap Gemini in Ireland decided it no longer wished to do ‘consultancy’ nor public sector projects. It made a number of us redundant and I then spent a number of years contracting.

I’m happy to say that, right up to the day I retired, I learned new things, which must be the secret to a happy career.

Last edit: April 2019

© John McEneaney 2019