Alan Turing: Codebreaker and Computer Pioneer
B.J. Copeland and Diane Proudfoot recall the contribution to the war effort in 1939-45 of the British computer scientist.
Before the start of war Turing turned to the problem of breaking Enigma, the coding machine adopted by the German navy in 1926, followed by the army in 1928 and the air force in 1935. The Enigma operator typed the plain message at the keyboard, and each time he pressed a key, a letter on a lampboard lit up. Each time the operator pressed a key, the internal wiring of the machine altered, thanks to three rotating wheels at the heart of the machine; so that if the operator repeatedly keyed A (for example) a succession of different letters would light up. At the receiving end, the cipher text was typed into a machine set up in exactly the same way as the sender’s, and the letters of the plain text lit at the lampboard.
In 1932 the Polish Cipher Bureau broke Enigma. But in 1939 the Germans increased the use of a plugboard which greatly increased the complexity of the machine. In July 1939 the Poles invited members of the British and French intelligence services to Warsaw, and ‘told everything that we knew and showed everything that we had’. For several months the British and the Poles collaborated. The latter recalled: ‘We treated [Turing] as a younger colleague who had specialized in mathematical logic and was just starting out in cryptology’. Yet he had already devised how to deal with the plugboard.
Turing’s ‘bombe’, or decoding machine, attacked the message text by means of a ‘crib’ – words likely to occur in the message. Cribs resulted from the stereotyped nature of the messages and the insecure habits of some operators. Weather stations regularly sent messages beginning ‘WETTER FUER DIE NACHT’ (‘Weather for the night’). In its mature form, the bombe contained thirty-six replica Enigma machines, with ten miles of wire and a million soldered connections. The prototype, named ‘Victory’, was installed in March 1940 at Britain’s codebreaking HQ, Bletchley Park (BP). By November 1941 there were fifteen bombes: at the end of the war there were several hundred. 1942 saw BP decoding 39,000 Enigma messages each month, and 84,000 a month by the autumn of 1943.
During 1940 BP read German air force traffic in large quantities, but naval traffic – including messages to and from the wolf-packs of U-boats in the North Atlantic – remained cloaked. If this traffic could be broken, the convoys could be routed around the the wolf-packs. When Turing took up residence at BP in September 1939, Naval Enigma was considered unbreakable. But Turing got interested in this problem, thinking that, as no one else was doing anything about it, he could have it to himself.
The problem was difficult because the sender enciphered the message setting – the trio of letters denoting the positions of the Enigma machine’s three wheels at the start of the message – by two different methods: once by means of the Enigma machine itself, as usual, and once by hand, using a table of letter-pairings issued to the operators. Called ‘bigram tables’ at BP, these changed on a regular basis. By the end of 1939 Turing figured out how this system worked, but the discovery could not be used until the bigram tables were known. Here the codebreakers depended on the Royal Navy. Several ‘pinches’ of material from German vessels enabled Turing’s ‘Hut 8’ to gain control of the code and, during June 1941, when the traffic was read currently for the first time, reroutings based on Hut 8 decrypts were so successful that the North Atlantic U-boats did not sight a single convoy for the first twenty-three days of the month.
During 1942 Turing worked on a new problem: ‘Tunny’, the machine used by the German army to encrypt teleprinter messages. Tunny messages were first intercepted in June 1941. Tunny was used for high-level signals such as messages from Hitler and the German High Command. It was not until mid-1942 that up-to-date Tunny traffic was read, using a method invented by Turing. Building on this, other members of BP’s Research Section designed the methods of attacking Tunny that were implemented in Colossus, the first electronic computer. Designed by engineer Thomas Flowers, Colossus was used from February 1944 to read the Tunny traffic. The timing of the D-Day landings was based on intelligence produced by Colossus.
Yet Colossus was not like a modern computer: to set the machine up for a new job, it was necessary to change some of its wiring by means of switches and plugs. The key idea of storing programs of instructions in memory, now very familiar to us, was Turing’s. Flowers said that, once Colossus had proved that large-scale digital electronic computing was feasible, Turing just had to wait for the opportunity to put his idea into practice. In 1945 Turing drew up a design for an electronic stored-program computer, to be called the ‘Automatic Computing Engine’.
Others who knew about Turing’s 1936 research also made the connection between electronics and the stored-program idea. A post-war race between Turing’s group in London and a Manchester group led by his friend Max Newman to build a stored-program electronic computer ended in June 1948, when the prototype electronic stored-program computer built by Newman’s group ran its first program. It was the start of a new era.
- Middle East
- North America
- South America
- Central America
- Early Modern
- 20th Century
- Economic History
- Environmental History
- Food & Drink
- Historical Memory
- Science & Technology