Alan Turing and the Ace computer

By Mark Ward Technology correspondent, BBC News

The BBC News series on British computer pioneers and pioneering British computers continues with the story of the Ace computer, which brought together a team who would go on to design the technology that underpins the internet.

Towards the end of the war, Alan Turing - the father of the computing age - had hid himself away in a hut at Hanslope Park in rural Buckinghamshire where, he told his assistant, he was "building a brain".

At the end of fighting, Turing took his plans with him to his new post at the National Physical Laboratory (NPL) in Teddington.

In March 1946 he handed over a report (which went unpublished during his lifetime) which contained detailed plans, including circuit diagrams, for the Automatic Computing Engine (Ace).

But when the engineers and scientists at NPL saw the plans they blanched at its complicated design.

Instead of building the whole thing, they decided to put together a smaller pilot machine. By this time, Turing had left NPL for a sabbatical at Cambridge and it fell to Jim Wilkinson, Harry Huskey and, later on, Donald Davies to get on with the construction.

The machine ran for the first time on 10 May 1950. By modern standards it was sluggish but in its day was the fastest in the world.

Error correction

Turing's vision for Ace was that it would complete entire calculations for scientists and researchers, rather than do the bits and bobs of mathematical jobs that computers typically did before Ace came along.

This made programming Ace a formidable task.

And, whilst investigating how it could be used, the team uncovered another problem that looked set to dog greater use of computers - how accurate were they?

"When you put decimal numbers in a computer they have to be converted to binary," said Professor Maurice Cox, who worked with Jim Wilkinson. "The conversion is not exact."

Binary is method of representing numbers using only the digits 0 and 1, used by all modern computers.

PILOT ACE Tech specs: 800 valves, 2816 bit memory, 15 instructions First operation: 10 May 1950 Creators: Alan Turing, Jim Wilkinson, Harry Huskey

"Errors in various stages of the calculation due to imperfections in the data can build up," said Professor Cox. "Those errors can explode if you have an unstable method of calculation."

Jim Wilkinson took on and defeated that uncertainty. Remembered with affection by everyone that worked with him, his work has been overshadowed by Turing.

"He was brilliant in his own right," said Clive Hall, a former colleague of Mr Wilkinson and who oversees some of the computer archives at NPL. "The problem was that he came to NPL when Alan Turing was there."

Wilkinson produced algorithms that could demonstrate the accuracy of computer calculations.

"It is work that became vital to all engineering and scientific calculation," said Prof Cox. It still is through the Numerical Algorithms Group, which produces libraries of algorithms used in used in diverse applications across the globe.

Network man

Another NPL pioneer, Donald Davies, also cut his teeth on the Ace. He joined NPL at the same time as Jim Wilkinson and was, for a while, Turing's assistant.

Much later, when he was head of the computer section at NPL, he did ground-breaking work on the best way to organise computer networks.

At the time making a phone call meant literally creating an electrical circuit between the two people in the conversation. That tied up the entire line for the length of that chat, even though for most of the time the connection will go unused because of the silences and gaps that punctuate conversation.

Rather than mimic this and tie up computer links for a long time as data was sent back and forth, Mr Davies realised that the spaces could be used.

By splitting data into packets and threading them on the same line, the carrying capacity of that link could be boosted and the whole network made more powerful.

Roger Scantlebury, who worked with Dr Davies, presented the ideas about "packet switching" to a conference in the US, where they were picked up by the creators of the nascent Arpanet, the fledgling internet.

Does that mean Britain invented the internet?

"Yes and no," said Mr Scantlebury. "Certainly the underlying technology of the internet, which is packet switching, we did invent."

The NPL network ran at multi-megabit speeds in the late 1960s, faster than any network at the time. The network was not just an academic toy either. Real work was done across it.

Like many contemporary machines, the full-size Ace filled rooms

David Yates was project manager of a program called Scrapbook which rolled together word processing, e-mail and hypertext - a system that anticipated many elements of the World Wide Web.

Scrapbook went live on in mid-1971 and became something of a "minor cult" among its growing user base, said Dr Yates.

"We had a community of bright people that were interested in new things," said DMr Yates. "They were good fodder for a system like Scrapbook."

Scrapbook helped people across the 28 acres of the NPL campus and beyond collaborate on projects without having to sit next to each other.

"When we had more than one Scrapbook system, hyperlinks could go across the network without the user knowing what was happening," he said.

"I cannot make claims for precedence for it," said Dr Yates. "But it is certainly the first system I know of to combine screen-based word processing, hypertext and e-mail as a service over a general-purpose computer network."

Whilst it is stretching the truth to say that Britain did invent the internet, there is no doubt that the history of the network is more complex than anyone ever thought.

E-mail this to a friend

Printable version