74 | 8 TURING AND THE ORIGINS Of DIGITAl COmPUTERS
In my view, Lady Lovelace’s 1843 account of programming (which even includes an allusion
to what we would now call indexed addressing) also shows a remarkable understanding of the
concept of a program, although the question of the extent to which she, rather than Babbage,
was responsible for the contents of her notes on this topic is not at all clear.
In fact, apart from these vague statements by Charles Babbage and Lady Lovelace, and of
course the very clear implications of Turing’s 1936 paper (with its infinite tape ‘memory’), the
earliest suggestion I knew of that instructions be treated as information to be stored in the main
computer memory was contained in von Neumann’s famous EDVAC report.^25 However, I then
found a 1945 report by J. Presper Eckert^26 containing the claims that in early 1944, prior to von
Neumann’s association with the EDVAC project, they had designed a ‘magnetic calculating
machine’ in which the program would ‘be stored in exactly the same sort of memory device as
that used for numbers’. Regrettably, there was no consensus regarding the relative contributions
of Eckert, Mauchly, von Neumann, and Goldstine to the design of EDVAC—a controversy that
I did not wish to enter into.
At this stage, the initial major goals of my investigation, in particular that of establishing
whether Turing had played a direct role in the development of the concept of the stored-
program computer, had not been achieved, though I did feel that my investigations had helped
to clear up some of the more important misconceptions and misattributions.
But, unfortunately, my investigation in 1972 of Turing’s post-war work at the NPL did not
match the thoroughness with which Carpenter and Doran later analysed Turing’s 1945 design
for ACE.^27 Rather, in 1972, I read too much into the fact that Turing’s report^28 on his design
slightly post-dated (and indeed cited) the EDVAC report,^29 whereas Carpenter and Doran
provided a detailed comparison of the fully developed stored-program facilities that Turing
proposed for the ACE, against what they showed were the rather rudimentary ones described in
the EDVAC report. (In contrast to Turing’s ACE report, the EDVAC report explicitly disallowed
the modification of stored-program instructions, although this restriction was later lifted.)
In retrospect it is clear that I really should have studied the EDVAC and ACE reports more
thoroughly, and perhaps assigned more credit to Turing regarding the origins of the stored-
program concept beyond that due to his 1936 paper. Clearly, I should later have included at least
some of his unfortunately little-known 1945 ACE report in my collection of selected papers on
the origins of computers. It was to be many years before I provided a much fuller analysis of the
stored-program concept and of Turing’s role in its development.^30
In fact, the question of who first had the idea (and understood its fundamental importance)
of having an extensive addressable internal memory used for both instructions and numerical
qualities, and also the ability to program the generation and modification of stored instruc-
tions, continues to be a matter of debate. What is indisputable, however, is that the various
papers and reports emanating from the EDVAC group from 1945 onwards were a source of
inspiration to computer designers in many different countries, and played a vital part in the
rapid development of the modern computer.
Concluding remarks
Although my initial goals had not been achieved I had—perhaps more importantly—managed
to accumulate some evidence that in 1943, a couple of years before ENIAC (hitherto generally
accepted as the world’s first electronic digital computer) became operational, a group of people