CAmPBEll-kElly | 217
reacted negatively to this suggestion, no doubt supposing that it would delay the building of a
full-scale machine. He took no part in it, left the NPL for a sabbatical year at King’s College,
Cambridge, in September 1947, and never returned. At the instigation of Charles Darwin,
director of the NPL, construction was handed over to the Radio Division, with the Mathematics
Division acting as consultants. Unfortunately, the Radio Division had no experience in digital
electronics, and it was not until the spring of 1948 that they established an electronics section
with some expertise. Meanwhile, by the end of 1947, Huskey’s year at the NPL was over and the
project was stalled.
The impasse was broken when members of the ACE section negotiated with the head of
the electronics section, F. M. Colebrook, and an active collaboration was established between
the two groups to build what had by then become known as the ‘Pilot ACE’ (Fig. 21.1). From
this point on, progress was rapid and the Pilot ACE sprang into life in May 1950. The machine
was equipped with thirty-two delay lines, each holding thirty-two words, giving a total stor-
age capacity of 1024 words—which was comparable to that of the Cambridge and Manchester
machines.
Optimum programming
From the moment that it began operation it was clear that the Pilot ACE was more than a small-
scale prototype. It was potentially a highly effective computing machine in its own right, with a
theoretical maximum speed of 16,000 operations per second.
figure 21.1 The Pilot ACE in the Mathematics Division of London’s National Physical Laboratory.
© Crown Copyright and reproduced with permission of the National Physical Laboratory.