I would say that fair play must be given to the machine. Instead of it sometimes giving no answer we could arrange that it gives occasional wrong answers. But the human mathematician would likewise make blunders when trying out new techniques. It is easy for us to regard these blunders as not counting and give him another chance, but the machine would probably be allowed no mercy. In other words then, if a machine is expected to be infallible, it cannot also be intelligent.
If Turing did not yet understand what it meant to be shown “no mercy,” he would, alas, learn that lesson all too soon. For the present, he was content to make a “plea” that his machines be treated with greater tolerance than he, as a homosexual man, was destined to experience:
To continue my plea for “fair play for the machines” when testing their I.Q. A human mathematician has always undergone an extensive training. This training may be regarded as not unlike putting instruction tables into a machine. One must therefore not expect a machine to do a very great deal of building up of instruction tables on its own. No man adds very much to the body of knowledge, why should we expect more of a machine? Putting the same point differently, the machine must be allowed to have contact with human beings in order that it may adapt itself to their standards.
It all came down to loneliness. Maurice Pryce, who knew better than Turing “the right things to do,” was getting on in his career. Other friends were marrying, having children, while Turing had become very much the “confirmed solitary” Newman had feared he might turn into. Now it seemed as if he was determined that the ACE—which had not yet even been born—should have a very different adolescence from his own; that it should enjoy “human contact,” and not be subjected to prejudicial injustices. There is in the lecture a clear sense of identification with the machine, as well as a certain protective affection for it—as if, having never found a life’s companion, the engineer in him was now determined to build one.
3.
Turing’s speculative musings on the ACE—in particular on its prospects for the future—sometimes reveal the same edge of paranoia that marked much of his odder behavior at Bletchley. For instance, late in the lecture before the London Mathematical Society, he divided those whom he saw as being destined to “work in connection with the ACE” into masters and servants, the masters being the theorists who decide its uses, the servants the technicians who “feed it with cards as it calls for them,” keep it in good working order, and assemble its data. As time goes on, however, “the calculator itself,” he hoped, could “take over the functions both of masters and servants,” with the latter in particular being “replaced by mechanical and electrical limbs and sense organs.” The risk, Turing imagined, was that the machine, by virtue of its very capacities, should come to pose a threat to human beings, who would in turn enter into a conspiracy against it (and perhaps, by association, its inventor)—a scenario not unlike the one that befalls Sydney Stratton in The Man in the White Suit. But the machine, so far, existed only in the form of a report: in prophesying such a dark and “dangerous” future for it, Turing was not just jumping past its construction, its testing, and its installation; he was taking its total success for granted. He was also putting the possibility of its dangerousness into the heads of the very listeners whose anxieties he was supposed to be trying to assuage.
The problem was not that he lacked the technical know-how necessary to design a computer; on the contrary, the report puts forward an astonishingly detailed battery of specs for the ACE’s design. The problem was that, lacking Maurice Pryce’s sense of social savoir faire, Turing was inclined to undermine his own chances of winning support by letting his imaginative flights of fancy get the better of him. As always, he said exactly what he believed—and suffered the consequences. For instance, in his lecture before the London Mathematical Society, he argued that it was much more crucial that the ACE be digital than that it be electronic. “That it is electronic is certainly important,” he told his listeners, “because these machines owe their high speed to this, and without the speed it is doubtful if financial support for their construction would be forthcoming. But this is virtually all there is to be said on that subject.” Speed, of course, had been the ENIAC’s raison d’être, while the ACE, with a pulse rate of a million a second, promised to be the fastest machine ever built. Yet, if Turing was to be believed, he was making the ACE fast only to appease the money bags on whose generosity its development depended. That he should have adopted such a patronizing tone not just toward the very people he needed to make the ACE a reality but to the ideal of speed itself was all the more bizarre, given the years he had spent trying to outrace the Germans at Bletchley. No experience could have made Turing more aware of the real value of a fast machine than the long days he had spent laboring to crack naval Enigma messages in time to prevent U-boat attacks.
Both the ENIAC and the EDVAC were also digital machines. The ENIAC, however, performed its calculations using decimal as opposed to binary notation, which bogged down the process, while the EDVAC put its emphasis on number crunching.* Since Turing, by contrast, imagined the ACE being put to many uses that did not involve number crunching, he designed it to be rather minimalist, with the emphasis on programming: in modern parlance, on software rather than hardware. Indeed, in his report Turing listed, among the many possible tasks for which the ACE might be employed, one of interest to the military (“Construction of range tables”); one of interest to pure mathematicians (“Given two matrices whose coefficients are polynomials of degree less than 10, the machine could multiply the matrices together, giving a result which is another matrix also having polynomial coefficients”); one of interest to engineers (“Given a complicated electrical circuit and the characteristics of its components, the response to given input signals could be calculated”); and one of interest to municipal governments (“To count the number of butchers due to be demobilised in June 1946 from cards prepared from the army records”). He also listed a function of interest to children (doing a jigsaw puzzle) and one of interest to himself (playing chess). The ACE’s digital nature meant that it could “be made to do any job that could be done by a human computer, and . . . in one ten-thousandth of the time.” And this was to a great extent because the machine was so remarkably simple, employing only the most basic symbolic vocabulary: “To perform the various logical operations digit by digit, it will be sufficient to do ‘and,’ ‘or,’ ‘not,’ ‘if and only if,’ ‘never’ (in symbols A & B, A ∨ B, ~A, A ≡ B, F).” More complicated arithmetic—even addition and subtraction—would be part of the programming, which again put the machine at odds with the EDVAC, in which the arithmetic was to be performed by feeding numbers into the machine’s accumulator.
Its digitalness, however, was not enough to guarantee the ACE’s success. The implementation of an efficient storage system was also central. In the lecture, Turing listed the three types of storage that in his view would best serve a machine such as the ACE: magnetic wire, cathode-ray tubes, and acoustic delay lines. Cathode-ray tubes, or even iconoscopes of the sort used in televisions, were, he felt, “much the most hopeful scheme, for economy combined with speed.” But they were not yet widely available in England, so he opted for the delay lines, which guaranteed the ACE the capacious memory that it would need if it was going to be independent. The ultimate point of having a big memory was to allow the computer’s operators—the servants—to forget the more tedious aspects of programming, which the machine would take care of by itself.
Clarity and concision were of paramount importance for the servants, who would presumably work the machine in ignorance of its engineering: “It should be possible to describe the instructions to the operator in ordinary language,” Turing wrote, “within the space of an ordinary novel.” And the language, for the sake of the servants, was ordinary, if a little morbid. One put away subsidiary tables, for instance, by “burying” them, a task one achieved through the use of “a standard instruction table BURY”; likewise, one fetched
the tables by “disinterring” them through the use of “the table UNBURY.” Not that the servants could create these ur-tables; instead, they would “have to be made up by mathematicians with computing experience and perhaps a certain puzzle-solving ability. . . . This process of constructing instruction tables should be very fascinating. There need be no real danger of it ever becoming a drudge, for any processes that are quite mechanical may be turned over to the machine itself.”
It was a scenario not unlike the one at Bletchley, where masters who had won crossword puzzle contests worked out the theory in one hut, while in another, Wren-like servants performed the day-to-day chores involved in running, and looking after, the beloved child: the machine.* What remained unspecified was the role that Turing—as inventor, father, and lover—was expected to play.
4.
Turing finished the report on the ACE in 1945 and gave it to Womersley, who in turn brought it before the executive committee of the National Physics Laboratory (or NPL) on March 19, 1946. The ACE’s creator also spoke at the meeting. His presentation did not go over terribly well—Darwin in particular appeared not to “grasp the principle of universality,” while Turing lost many members of the committee by letting his talk become overly technical. Nonetheless, in the end Darwin recommended that Turing be allocated £10,000 for the construction of a smaller version of the ACE—a “Pilot ACE.” Had he won the committee’s full confidence, he probably would have gotten more money, but a Pilot ACE was better than no ACE, and he accordingly set up shop in Teddington and went to work.
It was a transitional moment not just for him but for England. Yes, the war was over, but what was to happen next? And what would the future hold for Alan Turing, whose huge contribution to the effort at Bletchley remained (and would presumably remain for years) shrouded in official secrecy? Few at Teddington had a clue as to how much they owed him—a state of affairs that only intensified his sense of leading a solitary, ciphered existence. At Bletchley he had taken to long-distance running. Now he joined the Walton Athletic Club, the membership of which, Mrs. Turing noted in her memoir, “comprised men from all walks of life—road-sweepers, clergymen’s sons, dentists, clerks and so forth—he was always at ease among them and made them feel at ease.” By way of practice, he frequently ran the eighteen miles from Hampton-on-Thames, where he lived in a guesthouse called the Ivy House, to his mother’s house in Guildford. Likewise, when he needed to go to the Post Office laboratories at Dollis Hill, he ran the fourteen miles, usually wearing old flannel trousers tied at the waist with rope.
Perhaps thanks to its independence from the war effort, the atmosphere at Teddington was more bureaucratic and less encouraging of collaboration than at Bletchley, with a clear division between the engineers and the theorists. Turing was expected to function as an idea generator and leave the building to the engineers. At first the NPL made a point of publicizing its support for him. In an interview with the BBC, Darwin cast Turing as a sort of boy wonder (this even though he was already in his midthirties), explaining that “about twelve years ago a young Cambridge mathematician, by name Turing, wrote a paper which appeared in one of the mathematical journals, in which he worked out by strict logical principles how far a machine could be imagined which would imitate processes of thought.” From this arcane publication, Darwin implied, the possibility had dawned of a technological miracle from which ordinary Englishmen would benefit. Not surprisingly, such a mythology appealed immensely to the popular press, in particular the tabloids, which soon began calling Turing for interviews. His mother recalled one “evening paper” going so far “as to head a short paragraph about Alan with the words ‘Electronic Athlete.’ ” The emphasis was usually on the astonishing “feats” that the new “electronic brain” would be able to perform, in particular feats of memory to match those of any music hall memory whiz. For instance, the Surrey Comet quoted Turing as saying that the ACE would “be able quite easily to remember about ten pages from a novel, though not, of course, in their ordinary form. They would have to be translated into a medium it is capable of ‘understanding,’ in other words into the digits that it is designed to handle.” Similarly the ACE would at least in theory be able to play an average game of chess, though whether it could ever develop the “power of judgment” needed to play a good game of chess remained “a matter for the philosopher rather than the scientist.”
Though the “electronic brain” might be the darling of the Daily Telegraph, at the NPL it was beginning to cause some worry. Since news of the ENIAC had broken, there had been extensive information sharing between the Americans and the British. Womersley traveled to the States, as did Turing himself. Moreover, von Neumann’s influence had started to make itself felt at Teddington: if the EDVAC represented the direction that computer research was bound to take, would the British be foolish to follow Turing’s hunch and plan for such a different kind of machine? Would they end up being left behind? Or should the NPL hedge its bets? Notably, Maurice V. Wilkes, a former classmate of Turing’s and now the director of the Mathematical Laboratory at Cambridge, had attended a course of lectures in Philadelphia in the summer of 1946 sponsored by the group at the Moore School that had put together the ENIAC. Excited by what he had learned, Wilkes had returned eager to begin work on constructing a British version of the EDVAC. Although Wilkes intended the project to be based in Cambridge, he sought the cooperation of the NPL in putting together a plan for a machine that bore a much closer resemblance to the EDVAC than to the ACE. As if to underline the similarity, Wilkes’s computer was to be called the EDSAC—the electronic delay storage automatic computer—and it quickly won both the approbation and the attention of the NPL.* Turing’s ACE might fascinate the popular press, but it was out of line with the mainstream. In addition, its projected cost was skyrocketing.
In part, the problem, once again, was Turing’s deafness to the conventions. Although the press might cast him as a sort of Chatterton of the computer world—a “marvellous boy”—his peers knew better: proper Englishmen did not tie their trousers with rope. Nor did they jog to meetings in Dollis Hill. His insistence on going his own way was typical; now, however, he was asking England to take it on faith that it should follow him. And England balked.
At first it appeared that the NPL was going to give equal support to both projects—Wilkes, after all, had funding from Cambridge already—and when the report on the proposed EDSAC was completed, Womersley made a point of asking Turing to read it and give his opinion. Turing was not much impressed. “I have read Wilkes’ proposal for a pilot machine,” he wrote to Womersley on December 10, “and agree with him as regards the desirability of some such machine somewhere. . . . The ‘code’ which he suggests is however very contrary to the line of development here, and much more in the American tradition of solving one’s difficulties by means of much equipment rather than by thought.” No doubt Turing hoped that by appealing to the NPL’s nationalistic pride, he might ensure its continued support. Unfortunately, his tendency to run off at the mouth in interviews was proving to be a source of some embarrassment to the laboratory’s board. Womersley suggested that, rather than speaking with reporters, Turing should give a course of lectures on the ACE “intended primarily for those who will be concerned with the technical development of the machine.” Not surprisingly, Wilkes, who attended the lectures, complained that Turing was “very opinionated” and that his ideas “were widely at variance with . . . the main stream of computer development.”
They were—and remain so. Today the ancestry of most of the computers that we use can be traced back to the EDVAC—and not to the ACE, which in the end never even got built. Though Turing’s “minimalist ideas” were in Martin Davis’s words “destined to have little or no influence on computer development,” their legacy can still be felt in microprogramming, “which makes the most basic computer operations directly available to the programmer”; in the advent of the silicon microprocessor, which is in effect a universal machine on a chip
; and in the “so-called RISC (reduced instruction set computing) architecture,” which “uses a minimal instruction set on the chip, with needed functionality supplied by programming.” All of these owe much to the ACE.
The saddest part of the story, at least in Davis’s view, is the degree to which, for years, Turing got written out of the history of the discipline that he effectively invented. Though the Pilot ACE, for instance, survived, Turing had long since left Teddington by the time it actually got built, at which point it had gone through so many redesigns that it bore little resemblance to the machine of which he had dreamed. More cruelly, a 1949 report asserted that “the actual size of the ACE as originally contemplated was the outcome of long consideration by Mr Womersley and Professor von Neumann during Mr Womersley’s visit to the United States.” According to the evolutionary principles espoused by Sir Charles Darwin’s grandfather, it was perhaps inevitable that the soft-spoken Turing should end up being bulldozed by the debonair Johnny von Neumann. Indeed, as late as 1987, Davis reports, when he published an article claiming that von Neumann had gotten many of his ideas from Turing, he felt himself “to be very much alone” in that point of view. Davis was therefore gratified when twelve years later Time magazine not only named Turing one of the twenty greatest scientists of the twentieth century but in its entry on von Neumann (who also made the list) wrote,
The Man Who Knew Too Much: Alan Turing and the Invention of the Computer (Great Discoveries) Page 18