Prof

by Dermot Turing

  We are, Sir, your obedient servants

  A.M. Turing

  W.G. Welchman

  C.H.O’D. Alexander

  P.S. Milner-Barry

  Somehow Milner-Barry got past the policeman on the door despite having no ID, and handed the letter to a brigadier who did not sling him out on his ear. The letter was delivered to the great man. Churchill scribbled on a piece of notepaper and included one of his notorious stickers printed with the words Action This Day. His note read: ‘Secret. In a locked Box. Gen. Ismay. Make sure they have all they want on extreme priority and report to me that this has been done. WSC 22.x.’ None of the four conspirators – to become known as the Wicked Uncles – were fired, and the resources started to flow. There were also reorganisations of the Bletchley Park management; by February 1942 Denniston had been moved ‘sideways’ back to London to run diplomatic decryption, and Travis was left in sole charge at Bletchley.

  In the blackout

  Army and Air Force Enigma was producing copious amounts of intelligence. Bletchley Park was constantly trying to recruit more people, not just Wrens1 to operate the Bombes but professor types as well. Enigma keys changed at midnight, so a good deal of the decryption took place in unsocial hours. For most of the staff at Bletchley Park, shift work was particularly ghastly. Joan Clarke recalled later:

  Most people did not take their weekly leave when working the midnight-to-nine shift, but I can remember Alan Turing coming in as usual for a day’s leave, doing his own mathematical research at night, in the warmth and light of the office, without interrupting the routine of daytime sleep.

  The coming of the war hadn’t prevented Alan from continuing his work on formal mathematical logic. Throughout 1940 and 1941 he was corresponding with M.H.A. Newman on ordinal logics and the lambda-calculus. Writing from his digs at the Crown Inn, Shenley Brook End, Alan wrote periodically to Church with queries. ‘Dear Professor Church,’ he begins; Church responds, ‘Dear Dr Turing’. Alan’s relationship with Newman was warmer. ‘Dear Newman, Church’s notes certainly are rather a mouthful. I have never worked steadily through them myself, but have taken them in much the same spirit as you are doing. Fortunately I was able to go to the fountainhead for information.’ Together with Newman, still sitting out the war at Cambridge, he wrote a paper, which Newman sent off to Church for comments in January 1941; it was published in Church’s Journal of Symbolic Logic in March 1942. Another paper, this time by Alan alone, followed in the same journal at the end of the year. However, that was the last major contribution Alan was to make in symbolic logic. Alan’s intellectual interests were destined to take a different direction, because of a decision taken at the highest level in the German Naval Staff.

  Towards the end of 1941, hints had been dropped in the decrypted messages of a major change in German procedures. The old Enigma machines were going to be thrown out. Now the U-boats were going to use a four-wheel Enigma machine, and the codebreakers’ existing methods weren’t suitable for this horrific development. The switchover happened on 1 February 1942, and once again Bletchley was blacked out. The sinkings in the Atlantic started to rise once more.

  Improvements to the Bombe technology were needed. At the leisurely pace of 17,576 settings every 12 minutes, Bombes couldn’t find plausible settings quickly enough with Enigmas using an extra set of coding wheels: the increase to 456,976 possible wheel positions turned 12 minutes into over five hours, even if you could find a way to rig a three-wheel machine to solve a four-wheel problem. Six Bombes took 17 days to break naval Enigma in February 1942. Bletchley needed to add a superfast fourth wheel to its three-wheel Bombes. They turned to C.E. Wynn-Williams, who was working at the Telecommunications Research Establishment, for ideas; he was the expert in electronic counters, using gas-filled tubes which stood in for mechanical components and therefore worked much faster. These could monitor the positions of the fast wheel whizzing past each position in less than a thousandth of a second. Wynn-Williams designed a thing named the Cobra, so called because of its snake-like cable of 2,000 wires connecting it to the regular Bombe. The Cobra had all sorts of problems: one was the habit of its metal brushes bouncing and so missing fast-moving contacts; another was due to electromagnetic fields in the vicinity of the gas tubes, which made the switching unreliable. It was all quite difficult. Wynn-Williams’s high-speed Bombe was tried out but in October it was said that ‘the first Wynn Williams assembly has not been as successful as was hoped’. Doc Keen was working on a high-speed four-wheel machine too (‘it looks as if Keen’s machine will exceed expectations’), but under any circumstances production of fast Bombes was still several months away. Meanwhile the U-boats unerringly found their targets, and the British remained in the dark.

  A drawback of Alan’s regular three-wheel Bombe was that it only produced one plausible set-up for the three coding wheels, and one possible cross-plugging. There were nine more cross-pluggings to find, and then the ring-settings. All this could be done using the same crib and menu that had been used to wire up the Bombe, but the process involved a thing called a checking machine, which was operated manually. In theory the checking could be automated, and for this problem the Post Office Research Station at Dollis Hill were brought in to help. At Dollis Hill Alan Turing was introduced to another engineer, one who would bring Bletchley Park into the computer age before computers had even been invented. The man was called Tommy Flowers.

  Alan explained the Bombe-and-checking problem to Flowers. Explaining things was not always Alan’s forte, as Donald Michie described in a later interview with Professor Brian Randell:

  You had to get over a hurdle first before you were in the right ballpark at all, and it was very very easy for people to fail to clear that hurdle and consequently to regard Turing as totally incomprehensible, which was a very widespread attitude towards him, and this arose in two ways, one being simply overawed or panicked by a person of very great intellectual penetration. Secondly, his personal peculiarities were so obtrusive, as for instance his style of speaking and his whole appearance, and the way he twitched his head as he spoke and squeaked in a very high pitched grating voice and appeared to stammer, simply fazed some people.

  Although most people found Alan incomprehensible, this was not the case with Flowers, who ‘never had any trouble with him’. ‘I thought he was a charming chap once you got to know him, but he certainly was odd. Turing was always very much – complete in himself. You never felt that he was joining in, that he was dependent on anybody else, but he was a very charming chap.’ Despite their effective working relationship, the machine which Alan wanted was not a success: the speed specification was far too low for the task in hand, and the project was abandoned. There was a turf war, or a skirmish, between the British Tabulating Machine Company, who were building Bombes, and the Post Office Research Establishment, who were not. Notwithstanding, Flowers’s talents were not going to be thrown away – he had the impression that Bletchley Park felt guilty that his time had been wasted, and he had now been cleared for working on the most secret operations of the war. His would be a new codebreaking project, which in 1942 was just beginning at Bletchley Park, and it would involve Tommy Flowers and Alan Turing in the genesis of the first all-purpose computing machinery in the world.

  Meanwhile, there was a war to fight. On the intelligence front the fighting wasn’t going to plan, and turf skirmishes were breaking out in other theatres. Since December 1940 the British had been trying to arrange an intelligence partnership with the United States. The first unveiling of the Enigma secret took place in March 1941, but after then relations had deteriorated. Secrets are not shared willingly, and there was suspicion of non-disclosures and counter-suspicion of indiscreet sharing of secret material. Even when America entered the war it didn’t get much easier. Bletchley Park sent its head cryptographer, Colonel John Tiltman, to America in March 1942 to try to convince the Americans to focus on Japanese codes while the British handled Enigma. Given that Bletchley was in a b
lackout, the Americans weren’t in the mood to be convinced. They thought they should sort out naval Enigma for themselves, not surprisingly given that U-boats had been creating carnage with American shipping along the eastern seaboard of the United States. Eventually the British let them have answers to a detailed questionnaire, and a copy of Prof’s book. Blueprints for the Bombe were also promised, but those the British didn’t deliver; then in mid-1942 there was the Blue Bird Incident, involving Colonel B.F. Fellers, who was the US military attaché to the British Army in North Africa.

  On 24 April GC and CS had decyphered an appreciation from Kesselring1 referring to information he had received from a reliable source to the effect that a British attempt to advance to Benghazi was not possible before the beginning of June.

  ‘Reliable source’ is just how the British would describe the provenance of intelligence derived from cracking enciphered messages.

  Early on 24 June Sigint2 revealed that the Germans were aware, from decrypts of his signals, that in the opinion of the US Military Attaché in Cairo the British had been decisively beaten and that this was a suitable moment for Rommel to take the Delta.

  It was bad enough that the US attaché Colonel Fellers was a defeatist, but unforgivable that his signals were sent in a cipher broken by the Germans and revealed to Kesselring the Allies’ plans for the Battle of Egypt. Insult was added to injury because the revelation came to the British through their own codebreaking activities. Some serious bridge-building was needed to overcome these setbacks. Two American naval officers, Lt Cdr R.B. Ely and Lt (junior grade) Joe Eachus, were posted to Bletchley Park, arriving in July, and both worked closely with Alan Turing. The Blue Bird Incident was bad enough to implicate the Americans, but it was worse than that. The indications late in 1942 were that it was also British cipher security that was shaky. Although the British were blacked out of naval Enigma, they were reading Italian and Japanese ciphers – and these other decrypts gave the British the fright of their lives. The Axis powers seemed to be able to read the British Naval Cypher No. 3. Not only did the convoys have no idea where the U-boats were; the U-boats knew exactly where the convoys were. It was a disaster of the first order. The Allies needed to sort themselves out, and get back on top of naval Enigma.

  Commander Travis worked hard on the problem, and together with Commander Joseph Wenger of the US Navy, a deal was thrashed out. Among other things, the British were to provide ‘technical assistance in the development of analytical machinery’ and agreed ‘in principle to full collaboration upon the German submarine and naval cryptanalysis problems’; the British were ‘to obtain certain items of special analytical equipment developed by the U.S.’ The quid pro quo was that Britain would leave Japanese naval cryptanalysis to the Americans. To help with the ‘technical assistance’, Travis agreed that he should send Prof across to liaise with the US Navy Department:

  FOR ‘OP–2Ø–G’ FROM ‘G.C. & C.S.’ X T163

  FOR WENGER FROM TRAVIS

  YOUR WB17Ø X SHOULD BE GLAD IF TURING (WHO IS NOT A PROFESSOR) COULD COME EXAMINE MACHINERY X MAKE ANY USE YOU LIKE OF HIM IN CONNECTION WITH BOMBES X HAVE SUGGESTED HE STAY A WEEK IN WASHINGTON BUT IF YOU WOULD LIKE HIM LONGER I SHOULD BE QUITE WILLING

  Although the successes of the U-boats in the Atlantic were at their height in November 1942, Alan took a passage on the RMS Queen Elizabeth, whose average speed was, at 26 knots, about 8 knots faster than a surfaced U-boat. He arrived in New York for a third time on 13 November 1942. This time he wasn’t going for the attractions of Princeton in the fall. He had another mission, which had very little to do with breaking codes. Alan Turing’s new responsibility was to prevent the British making mistakes like those of Colonel Fellers.

  Notes

  1 ‘Daughters of two members of [Denniston’s] Ashtead golf club whom he knew well’, recruited in August 1939

  1 Michie joined Bletchley Park in 1942 and later became Professor of Machine Intelligence at Edinburgh University

  1 Plain language

  2 Peter Twinn, Dilly Knox’s chief assistant

  1 WRNS or Womens’ Royal Naval Service

  1 Commander-in-Chief of the German forces in North Africa

  2 Signals Intelligence

  7

  LOOKING GLASS WAR

  ALAN TURING’S TRIP to America was so secret that, nine years later, in writing to his friend Norman Routledge Alan said that the job he had had during the war ‘certainly did not involve any travelling’. In fact, his work between late 1942 and the end of the war remained under official wraps right into the twenty-first century. Even at the time of his departure for the United States in November 1942 it was so secret that he was told to take no documents at all other than the contents of a diplomatic bag; taking this admonition rather too literally Alan had a difficult time with the immigration authorities on arrival in New York.

  I reached New York on Friday November 12th. I was all but kept on Ellis Island by the Immigration Authorities who were very snooty about my carrying no orders and no evidence to connect me with the F.O.1 They considered my official’s passport insufficient in itself. They asked me very minute details about where I was to report etc. I think it might have been better from a security point of view if I had been provided with some kind of document of the kind they wanted.

  Alan stayed a few days in New York, where he met with a Canadian engineer called Lieutenant Colonel Pat Bayly. Bayly’s Number One project was to improve the security of communications across the Atlantic: between the British on both sides of the ocean, and between the British and the Canadians, all without being overheard by the Americans.

  Bayly’s focus was messages sent in the teleprinter code. Instead of using Morse, teleprinter messages go out in the five-bit binary Baudot-Murray code. As with all ciphers, if you add the key to the plain-text, you get gibberish; the clever part is that since the Baudot-Murray code is binary, adding the same key to the gibberish gives you back the plain-text. Bayly had invented a machine which generated a random string of binary digits, which were punched out onto two tapes: one which could be added by the cipher-clerk at the sending end, and another, containing the same key-string, which could be added by the cipher-clerk at the receiving end. Then the tapes would be destroyed, so that the key would never be used again. This method, which is the ‘one time pad’ system of encipherment, is the only truly secure one – assuming it is properly used. Bayly’s machine for producing the ‘tapes’ which carried the key-string was pronounced by Alan to be secure.

  By the end of 1942 – the turning point in the war described by Churchill as ‘the hinge of fate’ – Alan Turing’s job had itself turned about. His meeting with Pat Bayly was a foretaste of something which would occupy the next two years of his secret career. It was becoming the mirror image of its previous shape: Alan was changing from safe-breaker into safe-keeper. From this point on his cryptographic role would mainly be to guard secrets, rather than to uncover them.

  Dr A.M. Turing, Ph.D

  After Alan’s vetting of Bayly’s work, it was time to check up on, or, to put it diplomatically, to liaise with the Americans. It wasn’t apparent that the Americans wanted to be liaised with, or checked up on, which is probably why Alan had struggled to get off Ellis Island. The liaison mission was to begin with the American cryptanalysts of Op.20 G, the US Navy’s equivalent of Bletchley Park. After a few days with Bayly Alan moved on to Washington, and had a few awkward days there. First up was a meeting with the Americans’ answer to Dilly Knox.

  As I saw nobody working with pencil and paper, I asked if there was anyone in E1 who did so. Was introduced to Mrs. Driscoll at this point. I was rather alarmed.

  Alan might well be alarmed. Agnes Driscoll was a formidable intellect, who, like Dilly, believed in the power of the pencil. Her attack on naval Enigma had proceeded from first principles, and she believed it possible to find a ‘way in’ using an eight-letter crib. Unfortunately Alan had shown conclusively that it was in practice impossib
le to break three-wheel naval Enigma this way: Mrs Driscoll’s method would give about 3,000 solutions for each wheel order, with 336 wheel orders to try. He wrote a memo, which in true British fashion did not rubbish her approach, but it concluded with a set of unanswerable questions, suggesting only that the British must have misunderstood her method.2 At their actual meeting Mrs Driscoll tried the same technique herself, asking Alan ‘a great number of questions, to most of which fortunately I did not know the answers’, and the moment of danger passed.

  As requested by Travis, Alan was also in America to examine machinery. There were obstacles when foreigners wanted to look at secret machines, and Alan found that ‘my Princeton Ph.D. was quite useful in enlisting help, so I have decided to go by the name of Dr. Turing officially whilst I am over here’. His first report back to London was suitably doctored:

  MOST SECRET

  Washington

  November 28, 1942

  REPORT BY DR A.M. TURING, Ph. D.

  Alan was next due to visit Bell Laboratories, the research division of the telephone company AT&T. In many ways this was the counterpart to the Post Office Research Establishment where Tommy Flowers was based, and it was where they were building Bombes for the US Army. Getting into Bell Labs was difficult. For one thing the US Army was not the US Navy; Alan’s security clearance was a naval one, and so self-evidently invalid. So, for the time being, Alan went to see the Americans’ own four-wheel naval Bombe, which was not at Bell Labs. Although the US Navy reckoned they should be able to move ahead faster than the British with a four-wheel Bombe, it was not proving to be so easy. And the Americans were using an awful lot of equipment.