Software

Computers and software have been a large part of my working life

Fresh out of university in 1969 with a (not good) pharmacy degree, I decided to switch tack and got my first real job – in computing with Burroughs Machines. At that time business was transitioning from using electro-mechanical, to semi-electronic accounting machines. The first machine I worked on was the E2000 – it basically had a large carriage, bigger than a typewriter, which chugged around printing things on a paper roll. Each time it chugged, a set of tappets would fire against the underneath of the carriage at that position to read the program. The bottom of the carriage was actually a “magazine” – a flat panel into which I would have riveted pins of various sizes into various channels, and these pins would tell the machine what to add, subtract etc. and which way to move next. Thus the program was a mixture of position, direction and mechanical instruction. The next machine in the series, the E4000, actually had a special riveted pin which told the machine to execute an electronic program that I would have stored internally – that means in the sense that we now know a computer – though you can see from the picture below that this rudimentary micro-computer was the size of a massive cupboard.

5878367796_b05f25e33d_b

But the real joy was working on E6000s and E8000s – the programming of these was almost entirely electronic. They sported large cabinets, maybe two metres high and wide, and ametre deep, with a bank of red flashing lights on the front. (The world was still in black and white then.) And they could have awesome peripherals, such as punched-card readers, magnetic-card readers and line printers attached to them.

1029104_E_6000_Equipment_Reference_Manual_Nov68_0003

Looking back I cannot exaggerate the thrill I felt being around those machines with their flashing lights, and the excitement of writing the programs – I remember writing my first real program on Christmas morning 1969 – to work out a square root. The instructions were in machine language – there would be a number code for the type of operation – 20 was subtract, 58 was multiply, (yes – I can still remember the codes – things learnt young persist in memory) and another number code to indicate the memory address. I think we had 400 words of memory at our disposal – that is about 1K in modern terms, or 0.001 Megabytes!

In the 1970s I worked on a Digico computer, writing a set of business application in Basic for a clay quarrying company (WBB). The computer came with no database so I had to write a rather natty method for randomly accessing records which I called VEBAX (Virtual Expandable Bucket Access). After a few years the company upgraded to Hewlett Packard equipment, so we had to port the software and data across. I remember illicitly running a wire from one computer to the other, (and blanking out the bit of the operating system that crashed when it detected the connection) in order to get the information across automatically; then writing a translator (pretentiously called Rosetta) to convert the programs from one language to another. The only alternative in those days would have been to re-type it all – there were no memory sticks, or even floppy disks yet.

Back in the early 1980s, home computers were primitive, compared to today, but making inroads into homes, mainly for gaming. I teamed up with a colleague, David Styles, and under the name of Sector 7 Software, we released several pieces of software to run on the Oric and Atmos computers. You can see from the screenshots below that the available resolution was not great!

Untitled

I wrote what I believe was one of the very first music sequencers, released in 1983. Programmed in the ‘Forth’ language (does anyone still use that?), it allowed 3-part sequencing of music and some control over the envelope of the sounds. Software was supplied on cassette tape back then, and the cassette inlay was a concertina-folded 12-page instruction manual!

Untitle2d

There are several Oric emulators available running on today’s PCs, and so the games are still downloadable for free (thanks to archivists), and playable.

The Oric market was not lucrative however, so David and I switched to writing complex database re-formatting software (which we named Flexibase) to run on the Hewlett-Packard 3000 systems (which were prominent in business IT right up until the turn of the century). Our software successfully competed against a couple of American companies, and sold round the world.

Teaming up with another colleague, Mike Whiteley, we proceeded to code similar functionality for the next generation of  SQL databases, allowing a front-end PC to control the operations on Unix computers. In 1995 this product was acquired by Bradmark, a prominent software company in Texas, who proceeded to market, on-develop and support it, under their own name.

After a break for a few years, I got bored and decided to return to university to study biochemistry, having been inspired by a early book on the aspirations of genetic engineering. I followed on into researching for a PhD with wild ideas about DNA, but partly because my professor ran out of grant money, and partly because the protein that I was trying to work with was proving difficult to mutate, I gravitated back into writing software, this time to try to predict protein folding. At the time we thought the answer was writing more efficient processing algorithms, but since then I have come to believe that the physical descriptions of molecules are insufficiently accurate and incomplete.

Since it takes perhaps at least a whole day to read a novel, I was concerned to try to find some way of making sure that the commitment of choosing a novel was well-founded. In 2005 I developed a website based on the idea that millions of people would list their 10 or 20 favourite novels – so that then it would be possible for the website to do sophisticated matching of lists to come up with good suggestions for reads that would almost certainly be enjoyed. I never actually put that website on-line – I notice that the GoodReads.com website seems to do something similar, but they don’t make a big deal of it, so I’m not sure how sophisticated it is.