Patrick O’Beirne had already acquired an expertise in financial modelling when the first spreadsheet software appeared. For more than 40 years he has advised, trained and supported spreadsheet users. His company, Systems Modelling, also provided consulting services in software quality, application testing and spreadsheet risk.
He has served on the council of the Irish Computer Society and became a fellow of the association after leading its campaigns to prepare for the year 2000 date change and for the introduction of the euro currency.
My trajectory mirrored the development of the microcomputer business over 40 years. However, my first software experience was on an IBM mainframe.
UCD ran an IBM 1620 in 1966, when I was studying chemistry there, and I got access to the system. I thought that learning FORTRAN would be a help in drawing a titration curve showing how pH changes when two solutions are mixed. However, I didn’t think about scaling and used rather a lot of line-printer paper for a character graph, which drew a caution from the supervisor !
After graduation I joined the chemistry lab at NET, which made agricultural fertilisers in Arklow. My first work was routine analysis to support operations and later I did some research into new products. I learned to program a HP calculator at one point.
The accounting department at NET ran an NCR 500, but the company used bureau services for technical applications. The Arklow organisation developed its own software on IBM’s Call/360 timesharing service in the early 1970s.
One of my NET colleagues, an engineer named John Aherne, had a significant influence on my career. He had an MA degree in systems engineering from Lancaster University and, using the IBM bureau service, created NET’s first optimisation models. He inspired me to focus on computing. Andy Cusack, the engineering manager, supported this ambition and encouraged me to apply for the one-year course in Lancaster. This meant returning to full-time education, but he assured me that my future position would be guaranteed.
I took the masters degree in 1972/3. We studied statistics, systems thinking, FORTRAN programming, management science and operations research. There was also a summer placement and I was sent to Cadbury’s in York to work on a management science project there.
In or around 1974 the Arklow facility installed a Digital Equipment PDP-11 to extend its computing facilities to the engineering side of the business. Billy Lee was the operator of this system. He dealt with Digital, where Vernon Clay was our account manager. NET also hired two new programmers to work on payroll and accounting on the PDP-11. The corporate planning department, where I was based, was responsible for technical software.
Because of its mathematical characteristics, financial modelling was treated as a technical application. For example, NET wanted to use a linear programming technique for product mix optimisation. There was no packaged software for the PDP-11 that met this requirement, so I wrote one myself with Digital’s version of BASIC. I found BASIC to be quite a simple language and learned it by doing things. I started by copying projects from the DECUS catalogue and extended them as my grasp improved. I developed utilities such as a better text editor and a virtual array manager that evolved into a financial modelling package. I had found my niche.
In the mid 1970s NET decided to use natural gas from Kinsale Head to make ammonia at Marino Point in Cork. Our corporate planning department started working on financial models for this project. The finance specialists drew up a series of assumptions for a model and I was given responsibility for coding it. We used a dial-up link to a computer at IBM Ireland, then switched to a financial modelling package called Oracle (not related to the database company) that was run by Time Sharing Ltd (TSL) in London.
NET, however, was in trouble. Competition from cheap imports and changes in fertiliser production technology made the Arklow operation unprofitable and the Marino Point project was hit by escalating costs. The company experienced a severe financial crisis in 1981 and shed 350 jobs.
Systems Modelling’s Management Oriented Decision Evaluation Language System (MODELS) ran on Digital’s VAX and PDP-11 platforms. Based on Patrick’s work on financial modelling at NET, the software could also support production scheduling, forecasting and simulation. The company produced this brochure for the launch of MODELS at the Computex exhibition in 1983.
Cartoon by Martyn Turner. Brochure created by Bennis Design.
Shortly before that event I became aware of the emergence of microcomputers. My manager at NET, Peter O’Driscoll, told me that he had seen a demo of a calculator which automatically redid calculations as you entered data. That was VisiCalc on the Apple II which, at that time, was essentially the hardware that ran VisiCalc. I replicated that calculation feature in a grid view of our in-house financial modelling software.
In 1981 I left NET and set up my own business, at first as a sole trader and after a year as Systems Modelling, to sell the Apple II computer. I applied my modelling skills to VisiCalc and started training other people to design and run spreadsheets. The main incentive for them to buy a computer was that that it was faster than doing things by hand.
I joined the Irish Computer Society and started going to its events. Given my low membership number, this was possibly in the late 1970s or early 1980s. I also joined the IDPM in the UK and various other professional societies as my interests developed.
I also continued working on Digital minicomputers. I ran a training course for Alan Lush at Allied Computer Training and received a PDP-11/35 as payment in January 1983. This minicomputer enabled me to offer bespoke software development. Dennis Fitzgibbon, the TSL bureau’s agent in Dublin, gave me a start as a freelance BASIC developer on the PDP-11. After I stopped using the PDP-11/35 I donated it to the Irish Astronomical Society in UCD.
Systems Modelling started trading from an office in Arklow but later moved to Gorey. To illustrate how different the times were then, an invoice from the Department of Posts and Telegraphs states “Telephone service is usually provided within about two or three months of the agreement being signed”.
In the early days I employed one sales person, but soon found that the company did not really need one. I just got on and did my own thing without thinking too much about other dealers or how to differentiate my services from theirs. Like many entrepreneurs, I tried lots of ideas, such as a VAX & PDP-11 financial modelling package (MODELS) and attempted to export some industry-specific packages which I developed. Eventually I learned what worked well for me and what to let go.
Later, though, I wished that I had paid more attention to an IMI business course that NET had sent me on as part of the redundancy package.
This testimony continues in the ‘Lives of the PC dealers 1978-88’ archive. Click here to view.
Last edit: October 2022
© Patrick O’Beirne 2022