216 | 21 ACE
they could be placed so that they emerged from the delay line just at the moment they were
needed: this would eliminate the waiting time, making the machine potentially much faster.
To achieve this, each instruction would nominate the location of its successor. Numbers would
also be strategically placed to minimize waiting time. Although this made programming more
difficult, Turing was convinced that the superior performance would justify the added com-
plication: this coding style was later known as ‘optimum programming’. The idea of optimum
programming was wholly original and is perhaps the most direct connection with Turing’s 1936
paper. In both the Turing machine and ACE, each instruction nominated a successor.
from ACE to Pilot ACE
Following acceptance of Turing’s ACE report, Womersley explored the options for the
machine’s construction. These options included construction by the Post Office or collabo-
ration with the Telecommunications Research Establishment (TRE), both of which came
to nothing. In November 1946 he approached Wilkes at Cambridge. Wilkes, who had his
own funding uncertainties, responded positively by sending Womersley his initial plans for
EDSAC. Womersley passed these to Turing for expert comment. Turing’s memo—admittedly
a private communication—was nonetheless extraordinarily barbed and condescending. He
wrote:^4
I have read Wilkes’ proposal for a pilot machine, and agree with him as regards the desirability
of the construction of some such machine somewhere. I also agree with him as regards the
suitability of the number of delay lines he suggests. 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. I should imagine that to
put his code (which is advertised as ‘reduced to the simplest possible form’) into effect would
require a very much more complex control circuit than is proposed in our full-sized machine.
Womersley was a good deal more diplomatic than Turing in his response to Wilkes—who
was not shown Turing’s memo—but it was clearly not possible to bridge the gulf between
EDSAC and ACE and the two projects went their separate ways. Wilkes, in fact, did not see
Turing’s memo until 1977, when it was published by Woodger. In 1946 Wilkes had considered
that Turing’s ideas were wrong-headed, and in 1977 he believed that history was on his side.
During the period from December 1946 to February 1947 Turing gave a series of lectures on
computer design at the Adelphi offices of the Ministry of Supply in London (see Chapter 20).
Wilkes attended the first lecture, but no more—he thought that Turing’s ideas were so far
removed from the mainstream that there was no value for him in attending.
Meanwhile, construction of ACE was going nowhere fast. In January 1947, however, an
American named Harry Huskey arrived for a 1-year visiting position in the Mathematics
Division. He had worked on ENIAC and came to the NPL on the recommendation of
Douglas Hartree, Professor of Mathematical Physics at Cambridge University, one of the first
British experts to visit the Moore School after the war and a member of the NPL Executive
Committee.
Huskey made the suggestion that instead of going outside the NPL to get the machine con-
structed they should build it themselves. Well seasoned by his ENIAC experience, Huskey
advised the construction of a small-scale prototype which he called the ‘Test Assembly’. Turing