BEST OF BRITISH

Dick Pountain/PC Pro/Idealog 226/05/05/2013

A wave of nostalgia for the British home computer industry is upon us. It's mostly driven by games players who were schoolkids during that brief "golden half-decade" between 1980 and 1985, which is hardly that surprising given that so few of those UK-designed microcomputers were much use for anything else. Tony Smith has been running a highly entertaining series of memoirs about the Sinclair Spectrum, Lynx, Oric, Dragon, Jupiter Ace and more weird and wonderful devices at The Register website. Also emulators are available, written by selfless enthusiasts, to run all your old Spectrum games on a modern PC.

I'm an in-betweener, a crucial decade older than this Spectrum Kid demographic, so my own retro-spectacles are tinted rather less than rosy (indeed, closer to pale blue). To be sure I must thank the home computer boom for my present career, having entered the magazine business on the crest of it, but the machine I actually took home was a Sharp MZ80B running CP/M 3.4, on which I wrote my first book and learned Pascal, Forth and Lisp. To me computers were already serious tools rather than toys. In *my* schooldays I'd helped build an analog computer out of ex-RAF radar parts, and as a biochemistry student in the '60s I'd used London University's solitary Atlas mainframe to process my scintillation counter readings.

A month or so ago I had tea with Andy Hopper, Cambridge Professor of Computer Technology, President of the IET (Institution of Engineering and Technology) and for many years head of the Cambridge Computer Laboratory. We talked about Britain's role in the history of computing, and after our chat Professor Hopper sent me a copy of a lavishly illustrated new book celebrating the first 75 years of the Cambridge Computer Laboratory. I found myself quite transfixed by it, because although I sort-of-knew many of the facts it contains, I'd assembled the complete story of UK computing in my mind before, and the evocative B&W photos of the principal actors helped too.

The story of course starts in the 1840s with Charles Babbage's ill-fated attempts to build his mechanical Difference Engine - which now has a happy ending thanks to the superb working version now on display at the Science Museum. It carries on with Alan Turing, Bletchley Park, Colossus versus Enigma, a story now sufficiently familiar to make it onto TV and Hollywood movies. But there's another, less known story running in parallel with these landmarks. What's the working material that all modern computers manipulate? Electrons, as discovered by J.J. Thompson at the Cavendish Laboratory in 1897. In 1926, also at the Cavendish, brilliant radio engineer Charles Eryl Wynn-Williams invented a "scale-of-two" counter for an early radiation detector, which was the prototype of all digital devices.

The concept of the digital computer itself comes from Turing's 1936 paper on Computable Numbers. After WWII, with the triumphs of Bletchley still top-secret, British computer scientists found themselves in a neck-and-neck race with their US equivalents. Though Eckert and Mauchly's 1946 ENIAC is credited as the first working stored-program digital computer, the first practical one was EDSAC, built at the Cambridge computer lab by Maurice Wilkes's team in 1949. Made available to other university departments, its calculations contributed to several Nobel prizes including Richard Stone's for economics, John Kendrew's for the structure of myoglobin, and Martin Ryle's for radio astronomy. EDSAC's design also pioneered a bunch of crucial innovations still in use, including the subroutine, microcoded instructions and bit-sliced processor architectures.

But we lost the race because British scientists just don't have the Yanks' business acumen, right? Er, no. Wilkes was approached very early, in 1949, by the catering company J Lyons (of the Lyons Corner House cafe chain) to licence EDSAC as a model for the world's first commercially-useful business computer called LEO. Wilkes was quite adept at technology transfer and used Lyons' money to build EDSAC 2, while Lyons sold a range of three successive LEO models successfully until 1963, when taken over by English Electric (and later merged into ICL).

This story continues through the 1970s Cambridge Ring pioneering network project, which eventually lost out to Ethernet; to Acorn Computers and the BBC Micro; culminating with the formation of ARM Ltd as a joint venture with Apple (for the Newton PDA) which eventually saw ARM-designed CPUs driving the iPhone, iPad and most of the world's mobiles. We're often lectured nowadays that Britain's poor overall industrial performance is due to too much public and not enough private enterprise, but what this story says is that what our computing successes share with Silicon Valley's is that they're all started by scientists and engineers who understand the product (think Gordon Moore) rather than money men who don't.