By Mark Ward
Technology correspondent, BBC News
Bill Renwick and Maurice Wilkes brought Edsac into being
The BBC News series on British computer pioneers and pioneering British computers continues with a look at Edsac.
The Edsac computer is notable for many things including being the first computer designed for serious scientific use and for kicking off the use of machines in business.
But it must be unique amongst post-war machines in being designed onboard the Queen Mary.
Sir Maurice Wilkes started sketching out the design for what became Edsac on the cruise liner when returning from the US in 1946. He travelled to America to attend classes on how to build a computer given by Presper Eckert and John Mauchly.
Tech specs: 3000 valves, 4096 bit memory, 18 instructions
First operation: 6 May 1949
Creators: Maurice Wilkes, Bill Renwick, Tom Gold, David Wheeler, Eric Mutch
At that time they were the only ones widely known to have built such a machine. They called it the Eniac or Electronic Numerical Integrator And Computer.
Unveiled on 14 February 1946, Eniac had more than 17,000 valves, tipped the scales at 25 tonnes and consumed 150 kilowatts of power.
"It was an enormous brute, really," said Sir Maurice Wilkes. "All later computers were done on much more scientific lines."
It was on the return voyage that his ideas about how to build Edsac, the Electronic Delay Storage Automatic Calculator, crystallised.
"I knew exactly what to do," he told BBC News. "People like to make a song and dance about it but it was quite straightforward. With all my experience as an electrical engineer I just knew exactly what to do."
That hands-on experience had been gathered at the Telecommunications Research Establishment while helping to develop radar.
Following PhD work on the propagation of radio waves in the ionosphere, Sir Maurice had been invited to a training course at TRE. War broke out while he was there and a supposed five week stay turned into six years of hectic R&D.
Returning to Cambridge he was the perfect man to take over the mathematical laboratory and create a computer that could help the university's scientists carry out research.
Edsac was built to be used by scientists
Some of the hardware created for radar, in particular using tubes of mercury as a data store, were recycled for Edsac.
Beyond the hardware, Sir Maurice had very definite ideas about the uses to which the computer should be put.
Not for him the building of a computer simply as an intellectual exercise. Instead, he said, he wanted something practical.
"We wanted something that worked well enough and long enough to be useful in Cambridge," he said. "I think we did rather well in that respect."
"We knew that if we could provide them with digital computers we would not need to do any more," he said. "It would just take fire."
Helping Sir Maurice to build the machine were Eric Mutch, David Wheeler, William Renwick and Tom Gold. "A wonderful team," he said. Gold in particular was useful because of his experience with the tanks of mercury used on radar.
Edsac first ran on a program on 6 May 1949 and afterwards Sir Maurice is sure that the team celebrated.
"I'm quite sure we went to the pub," he said.
Cake and kudos
The machine was pioneering in that it was open to use by almost anyone that could put together a proposal to use it.
"Research students who were faced with several weeks hard work with a desk machine took to it very quickly," said Sir Maurice.
Edsac's memory involved long and short tanks filled with mercury
"Mostly the projects came from individual research students," he said. "Then they showed their results to their supervisors and the supervisors began to get the message."
The Edsac team maintained a library of sub-routines, chunks of code, that could be patched together to produce a wide variety of programs. They also gave students help on how to use the machine.
Two big projects were run on Edsac. One on radio astronomy and another on the structure of haemoglobin and myoglobin. For this John Kendrew and Max Perutz won a Nobel in 1962.
"The idea of computers spread from the bottom up which is not a bad way for it to happen."
Interest in greater use of the machine was not restricted to the students at Cambridge. Some of the money to develop Edsac came from cake and tea shop giant J Lyons.
It provided a grant and when the machine was finished decided it wanted one of its own. Lyons sent a team to the US to see Eckert and Mauchly to find out about computers.
Tea shop giant J Lyons was among the first to use computers
"They had a eureka moment that computers would be ideal for solving their problems," said Frank Land, a Lyons employee who become one of the first programmers for the machine the cake-maker went on to build. That machine, the Leo, was not just a straight copy of Edsac.
"They built a version of Edsac that had specific business features," he said.
"Edsac had very limited inputs and output equipment," said Dr Land. This had to be improved as Lyons wanted to use the machine to speed up its ordering systems - a procedure which involved a huge amount of data entry.
Once it got going, Leo was the engine that ran the whole Lyons business. Every day, said Dr Land, tea shop managers would ring head office and place their orders for the next day.
"Leo produced schedules for the factories to produce the stuff," said Dr Land. "It worked out the order they were going to be placed in the lorries and which cartons had to be used to pack them in."
Later on, he said, they experimented with time sharing to get more out of the machine and did work on using acoustic couplers so tea shops could automatically send back orders.