By Mark Ward
Technology correspondent, BBC News
Despite being a 'Baby', the Manchester machine filled a room
The BBC News series on British computer pioneers and pioneering British computers continues the story of the world's first programmable computer - Manchester's Baby.
The Second World War led to the creation of the code-cracking machine Colossus, and it played a similar role in the genesis of another pioneering computer, Manchester's Small Scale Experimental Machine (SSEM), known as Baby.
Rather than breaking German encryption, this one grew out of the work done on radar at the Telecommunications Research Establishment in Malvern.
One of the key developers of radar was Freddie Williams who, in 1946, became head of electro-technics at the University of Manchester.
Work on radar helped to bring the Manchester Baby to life
One project he took with him was how to use a cathode ray tube (CRT) to store data. In radar, the tube is used to display the position of any aircraft in the area of the sky that the system is scanning.
Travelling to Manchester were several TRE colleagues. Tom Kilburn had worked with Freddie Williams in the same room in the TRE cricket pavilion and Geoff Tootill ran technical support for airborne radar.
The pair, overseen by Williams, set about building a machine to prove that a CRT could be a reliable storage mechanism.
"They were absolute experts in using CRT technology," said Chris Burton, who ran a project to rebuild the SSEM or Baby.
"The tubes used in Baby were the commonest type that had been used in the war," said Mr Burton.
Those CRT tubes also had other virtues.
My assignment was to help construct a backing store for the next phase of development of a machine that Ferranti would take through to a commercial system. We were in competition with similar developments at the National Physical Laboratories at Teddington and the Cambridge Mathematical Laboratory.
The race to be first was on and done in the knowledge that we could cut corners while Ferranti engineers followed through with a more robustly engineered model of our prototypes.
Great fun, especially under the direct wing, in my case, of that arch-boffin Freddie Williams of what had by then become the Department of Electrical Engineering.
Years later a Ferranti manager of those times commented: 'They had all the joy of conception and none of the responsibility'.
We did not have a clue what would develop from that work. We only knew that we had to do it fast and, contrary to some of the eminent commentators of that time, with an intent to yield more than the three such computers which they said would be sufficient for the needs of the whole world.
"In terms of cost per bit, CRT memory is very cheap," said Mr Burton. "It does not require any precision equipment and uses only low cost components."
Tom Kilburn had proved his skills at TRE where he helped to design and produce circuits for use across the services.
Geoff Tootill had become a skilled engineer thanks to a childhood making audio amplifiers, long experience with radar and his abiding interest in back stage electronics for TRE's concert party.
His interest in hardware also meant he knew about the computational machines owned and operated by The Tote - the government body which handled pool betting at racecourses.
The Tote used binary as the basis of its computers and Tootill decided to do the same.
"At that time nobody had any experience of binary computers," said Mr Tootill. "There were desk machines for doing computation that all worked in decimals."
Tootill and Kilburn each designed the basic unit, the adder, of what would become the SSEM.
Tom Kilburn went for an analogue device and Geoff Tootill plumped for diodes and valves.
"Diodes cost less to buy and maintain," he said. "So reducing the number of valves meant it was going to be cheaper to build and more reliable in practice."
Freddie Williams had negotiated to ensure that the Manchester team could draw on the TRE stores for parts. That, said Mr Tootill, was a wise move.
"He knew that all the electrical components in retail shops and wholesale establishments was junk compared to the high quality ministry stuff," he said. "So he arranged a continuous supply."
Geoff Tootill's adder proved the more reliable and development of that went alongside refinement of the CRT storage device.
SMALL SCALE ELECTRONIC MACHINE
Tech specs: 500 valves, 1024 bit memory, 8 instructions
First operation: 21 June 1948
Creators: Freddie Williams, Tom Kilburn, Geoff Tootill, Tommy Thomas, Dai Edwards, Alec Robinson
"We were developing the CRT storage system and we were working in a fairly systematic way," said Mr Tootill. "We would design a new unit which was a chassis of particular dimensions carrying valves."
"We would draw the theoretical circuit diagram and we had a wire man who was skilled enough to translate that into the practical embodiment," he told BBC News.
"It was a steady progression," he said. "The reason we got on with it so fast was that we did not design anything that was not going to be used."
"I do not think I ever had to work as hard as I did then," he told BBC News. "We were working all day and at weekends too."
"Eventually we built up the equipment to the stage where it was running a program," he said. "It was a computer."
Chris Burton led a project to re-build the SSEM.
The machine first ran a program on 21 June 1948 and is recognised as being the first recognisable modern computer for two reasons.
Firstly, before the SSEM, computers were re-programmed by being re-wired. By contrast the SSEM used the store to hold both the data being worked on and the instructions about what to do with it.
Secondly, the SSEM had the first Random Access Memory (RAM) in the form of the CRT tubes. Modern computers typically sport a RAM several gigabytes in size compared to the SSEM's 1024 bits (128 bytes).
The CRT had the advantage that no matter which location data was stored in the memory, it could be read at the same speed. By contrast earlier machines used serial memory - typically sound waves passed through a tube of mercury - where memory could only be read in the same sequence as originally stored.
Celebration of that first successful run was muted, said Mr Tootill.
"We went to lunch in the canteen as usual and probably discussed what we were going to do next," he said. They decided to extend and expand the SSEM into the Manchester Mark I and cannibalised it for the parts to build the new machine. Although it disappeared, the SSEM has had a huge impact on the development of computers right through to the present day.