Early Years and the 6502
As a teenager I developed a strong interest in computing. I badgered my late father to buy me the Acorn Atom – the predecessor of the BBC micro. That was a major expense for him at the time and led to some significant sacrifices for him – something I absolutely did not fully appreciate at the the time and something for which I will be eternally grateful. I remember him spending all day on the bus to Cambridge and back just to get me some additional memory chips to expand the Acorn Atom. It’s been nearly 25 years since he died – still seems like yesterday. RIP Dad.
The Acorn Atom was a decent machine but not without it’s faults. It had a tendency to overheat – something that I solved through the judicious placement of ice cold cans of Coke above the heatsink. I taught myself a variety of programming languages on the Atom: Basic, Forth and 6502 assembler to name but a few.
After the Atom I dabbled in Z80 processor based machines like the Sinclair ZX-80/ZX-81 and Spectrum for a while, but one of my favourite machines around that time was the Oric Atmos – also a 6502 based machine. Mind you, loading programs from cassette tape on the Atmos was a bit of a disaster.
First Job, Z80 and CP/M
At work for the Department of Trade and Industry (DTI) at Warren Spring Laboratory in Stevenage in a non-computing scientific role, my 6502 programming skills were put to use on the Apple-IIe and BBC Micro computers that were used to collect data from hardware interfaces such as temperature probes. Although the BBC Micro was 6502 based, there was a Z80 co-processor add-on your could buy with a disk drive that allowed programs running under CP/M to be executed – something I had great fun programming. There was also a 6502 second processor which ran at higher clock speeds that we used for certain high performance data collection tasks.
British Telecom and the Motorola 68000
I always considered the Z80 processor less intuitive and less fun to program than the 6502 and everyone else seemed obsessed with playing games on those machines, rather than programming. But the 6502 was quite limited nonetheless.
Around that time I read a book on Motorola 68000 assembly language and I just knew I had to have a computer that used one. By that time I was working for British Telecom in Bedford, and one of my first purchases with money from that job was a Sinclair QL. Despite using the cheaper and less capable 68008 processor of the 68000 family, the Sinclair QL was truly a fabulous and innovative machine. It was so way ahead of it’s time but also limited by some poor design choices such as the ZX Microdrive.
My exposure to computing at work was very limited in my role in Bedford, so after a 18 months I applied for a computing role with British Telecom International in London. I passed the aptitude tests (they weren’t all that difficult quite frankly) and my “career” in computing began.
British Telecom International, COBOL and Mainframes
My first real computing job was all about mainframe computing, specifically COBOL programming on a IBM 3081 at the Barbican Computer Centre, just round the corner from my office in Farringdon Road. You only have to look at the specification of the IBM 3081 (up to 32 megabytes of RAM!) to appreciate how far the technology has advanced since then.
My major responsibility at BTI was bespoke COBOL programming and support and customisation of the McCormack and Dodge Millennium accounting software package.
Consulting, PeopleSoft and Linux
I left British Telecom International in April 1989 to join a small consultancy specialising in McCormack and Dodge software applications. This move gave me exposure to many different hardware platforms over the years including:
- IBM Compatible mainframes from Amdahl and Fujitsu
- ICL VME/B based mainframes
- DEC Vax VMS based servers
During this period, McCormack and Dodge was aquired by Dun and Bradstreet and became part of Dun and Bradstreet Software (DBS). DBS was subsequently bought by GEAC and then GEAC themselves were bought by Infor.
However, the move to client-server solutions was already well underway and spelt the death-knell for mainframe packages such as Millennium. Attempts at updating them for the new technologies or porting to smaller and cheaper hardware platforms were tried and failed. More agile software companies were taking over from the mainframe “big boys”.
My consultancy firm made the transition to PeopleSoft and became one of the first implementation partners in Europe in 1994. I transitioned into PeopleSoft over a number of years – learning it in-depth by effectively reverse-engineering “how it worked” from the COBOL source code that made up much of PeopleSoft applications at the time.
In 1993 I became a Director of the consultancy and as the company expanded I was exposed to a number of different technologies from a back-office perspective. We were early adopters of Internet technologies and were making use of Linux based e-mail servers running on 386 hardware as early as 1993. My Linux Journey is detailed here.
I continued working on PeopleSoft until the consultancy closed down in September 2001.
Since October 2001 I have been an independent contractor specialising in PeopleSoft applications. However, I have never considered myself limited to PeopleSoft – it is just an application suite and development environment and the technical skills I have gained over the years have applicability across a much wider sphere than just PeopleSoft.