Tommy's massive leap into dataage

When David Kahn's monumental history, The Codebreakers: The Story of Secret Writing , was published in 1967, a review in the Washington Post stated: "In importance, not to mention drama and sheer intellectual brilliance, the British effort pales every other account in Kahn's huge book, or perhaps all of them combined. That it has not been reported is less the fault of Kahn than of the British, whose still applicable legal strictures have maintained an astonishing public silence."

In 1975, after stories began to appear about work on electronic computing at Bletchley Park during the Second World War, and several books had been published containing information on the breaking of the German Enigma code there, the Government relented slightly. It issued photographs and a one-page description of a secret wartime computer called Colossus. I was then permitted to interview engineers and scientists involved with Colossus, and to publish a lengthy paper about it. By now, 30 years later, a great deal has been published on the work of Bletchley Park, and on Colossus - and a superb working replica has been created. Nevertheless, much vagueness and confusion has continued to surround the Colossus project, with many earlier writers wrongly assuming that it had to do with Enigma. Indeed, Paul Gannon, author of this latest book Colossus: Bletchley Park's Greatest Secret , claims that "it was precisely the intention of Britain's modern eavesdroppers that their predecessors' secret techniques and the real story of the Colossus should be confused with that of Enigma". In fact, Colossus was used against teleprinter traffic - typically between Hitler's high command and his army commanders - that was encrypted using a Lorenz SZ40/42, a cipher machine that is totally different from Enigma.

Gannon's book, aimed at a somewhat technically oriented readership, is one of the first to draw on the great mass of hitherto classified wartime documentation about Colossus that has recently been made available in the National Archives. The first eight chapters concern how teleprinter-generated radio traffic was intercepted. This was a very different problem from that of eavesdropping on Morse code and recording the Germans' Enigma-encrypted messages. Instead, it involved the development and use of special machines, called "undulators", which recorded the varying amplitude of the intercepted radio signal with a moving pen on a chart. These records could then be visually inspected and used to create a punched-paper teleprinter tape.

The next few chapters describe how Bletchley Park's cryptanalysts achieved the amazing feat of deducing the internal operation of a Lorenz machine solely through an analysis of two enciphered teleprinter messages, without ever seeing a machine. These two messages were discovered to be the result of a German cipher clerk typing in the same "plaintext" twice, almost but not quite identically. Bletchley Park's cryptanalysts went on to develop statistical methods for determining the machine settings and hence the plaintext of subsequent enciphered messages. These methods were impracticable without machine assistance - first, using mainly electromechanical machines, which processed information from two closely synchronised loops of punched tape, one of which represented the intercepted message. But large-scale continuing success awaited the invention of a much more complex machine by a brilliant engineer from the Post Office Research Station, T. H. (Tommy) Flowers. The aptly named Colossus employed an unprecedented number of electronic valves and dynamically generated the information that had to be compared repeatedly with the message tape loop, hence avoiding the difficult problem of reliably synchronising two high-speed paper tapes. Flowers had invented, and by late 1943 built, the world's first electronic computer, albeit a rather special-purpose one, controlled by plug-board and switches rather than a stored program.

What then developed at Bletchley Park was a truly massive code-breaking activity, employing ten Colossi by the time the war ended and hundreds of clerks, cryptanalysts, engineers and other personnel, an activity Gannon describes in great detail in the book's later chapters, and which he argues was even more important to the conduct of the war than Bletchley Park's work on Enigma.

Gannon's book contains a mass of utterly fascinating and largely unknown material about an immensely important wartime project, and as such is very welcome indeed. But it is not very well structured and, although the technical appendices are helpful, its lengthy descriptions of complex procedures and mechanisms are sometimes hard to follow. Thus the book would have been much improved if the author had received more editorial guidance.

This subject still merits the attentions of a writer such as Kahn.

Brian Randell is emeritus professor of computing science, Newcastle University.