Think about the history of computing, and the moments that spring to mind are likely to be Turing’s work cracking the Enigma machine, Babbage’s difference engine or Steve Jobs’ revolutionary work in Silicon Valley. But one major milestone in computing tends to be overlooked—and it took place right here in Manchester.
The Small Scale Experimental Machine, nicknamed 'Baby', was the world’s first stored-program computer. This computer was the first step to the world that we know today, as even modern computers build upon Baby’s basic principles.
So, who built this ground-breaking machine, and how did it change the world?
The birth of the Baby
Arguably, the modern computer age started in Manchester in 1948.
It was to Manchester that Sir Freddie Williams and some of his wartime team, notably Tom Kilburn and Geoff Tootill, relocated after their work perfecting radar during the Second World War. Williams had been appointed the head of the Electrical Engineering Department at the University of Manchester, which is where he earned his undergraduate in science before going on to get a PhD in Oxford.
Until this point, computers such as the codebreaking Colossus had to be physically reprogrammed and rewired every time they needed to do a new task. To take the first step towards the computers of today, which are capable of carrying out multiple tasks on command, researchers had to develop a computer with a memory.
To do this, the team at the University of Manchester developed the Williams-Kilburn tube, which used a cathode ray tube of the type commonly associated with old-fashioned bulky televisions. The 'screen' was a grid of dots, with each dot being the result of an electron hitting the phosphor screen of the tube. This creates a momentary charge that was used to write an operation into a computer memory. Originally, this was only a single bit—on its own not that useful, but the idea that it could be used to make a computer to remember a program was there. And that’s where Baby comes in.
Williams and his team needed to test their idea that the tube could create memory, so set about building a prototype stored-program computer. Using war surplus supplies, including some from Bletchley Park, Baby was built and completed by June 1948.
Once finished it was 17 feet in length by 7.4 feet tall, and weighed almost a ton. It seems huge, but the ENIAC computer built in America two years earlier—which couldn’t store a memory—was almost 95 feet long and weighed 27 tons. So, Baby was probably an appropriate nickname, even if we wouldn’t think so today.
On 21 June 1948, Baby ran the first of its programs, which was written by Kilburn and consisted of 17 instructions, that would find the highest factor of a given number. By August, Baby was successfully running the program with numbers as high as 2 to the power 18. To get to the correct answer, Baby ran through 3.5 million calculations in 53 minutes.
The answer is 131,072, by the way.
With Baby having proved that computers could remember a program, in August 1948 the construction of a more practical computer begun, which was known as the Manchester Mark 1. It was first operational by 1949, and it would directly lead into the development of the Ferranti Mark 1, which was the world’s first commercially available computer in 1951.
Related blog posts
What came next?
The Williams tube was used by a number of early computers on both sides of the Atlantic, and in 1953, under the supervision of Kilburn, Richard Grimsdale and Douglas Webb would demonstrate a prototype for a computer that used transistors, rather than the vacuum valves that had been used before this point. The transistors were cheaper than the valves, but to begin with, they certainly weren’t more reliable.
The original Baby had been built using army surplus supplies, and once it had been proven to work, it was disassembled and those part were re-used to develop the computers that came after. That meant that for 50 years, Baby didn’t exist. However, the task of building a working replica was undertaken in 1998 to mark the 50th anniversary of Baby’s first running a program by a group of volunteers from the Computer Conservation Society, who painstakingly rebuilt the machine from scratch.
As the original parts of Baby had vanished, the team—led by Chris Burton—tracked down the most authentic parts that they could find. This meant, for example, having to find Spitfire radio set switches, as the original had used them. In 1948 these had cost less than 20p, but by the time the re-build started in 1995 they cost £30.
Finding the old electronics that had been superseded decades ago was also a challenge, which resulted in going to collectors and dealers, but an even tougher problem then presented itself as the electronics had to be assembled so that they would work—but there were no circuit drawings of the original Baby. This led to the circuits being put together using photographs and the detailed drawings as Baby was being converted into the Manchester Mark 1.
Finally, the Post Office steel racks that held all the electronics in place also had to be located, with two of them being found in a garden in Shropshire—propping up a flower bed.
The original program that had been used had also been changed and modified after the first successful run, meaning that just a few months later the coding was different to the original. Therefore, for the replica to run the original program in its original state, Kilburn and Tootill went back over their notes and tried to remember their thinking at the time. While the original program was written on the train to Dewsbury, the recreation was a puzzle that was put back together 50 years later.
But by the time the working replica of Baby was finished, Tom Kilburn said that it looked identical to the original, other than being cleaner.
Watch volunteers run our replica Baby and see how far computing has come since 1948.