> I was surprised to read that the name ACE (automatic computing engine)
> was influenced by Babbage's work. I'd be interested in documentation.
> I was also surprised to learn that as many as 30 DEUCE's had been built.
> Turing's original ACE report is a gem that remained unknown except to a
> few for many years.
My sole source for both statements was Anthony Ralston _Encyclopedia of
Computer Science and Engineering, 2nd Edition_ (New York: Van Nostrand
Reinhold Company, 1983) article by J. H. Wilkinson on "Turing" page 1539:
"In 1945 [Turing joined] the newly formed Mathematics Division at the
National Physical Laboratory (NPL). His early work on computability,
combined with his awartime experience in electronics, had fired him with an
enthusiasm for working on the design of an electronic computer. The Machine
he designed, which was called the Automatic Computing Engine (ACE) in
recognition of Babbage's pioneering work, was characteristically original.
Although Turing knew something of the von Neumann proposals for EDVAC, he was
not unduly influenced by them. The ACE, as Turing conceived it, was too
ambitious a project, considering the current state of electronic techniques.
Therefore, he left NPL in 1948, dissatisfied with the rate of progress.
"...After Turing left NPL, it was decided to build a pilot model embodying
Turing's ideas (the Pilot ACE), and this was completed in 1950. It was a
highly successful computer and some 30 engineered versions of it were
subsequently constructed by the English Electric Company under the name
DEUCE. The original Pilot ACE is in the Science Museum in Kensington, London.
"...REFERENCES
"1955 Newman, M. H. A. _The Biographical Memoris of Fellows of the Royal
Society_ 1 London: the Royal Society, pp. 253-263
"1959 Turing, Sarah _A. M. Turing_ Cambridge: Heffer & Sons
"1970 Wilkinson, J. H. "Some Comments from a Numerical Analyst" (The 1970 A.
M. Turing Lecture) _JACM_ 18, No. 2: 137-147"
Now that you mention it, the figure of 30 DEUCES does seem improbable.
Thomas Watson Jr. (page 242) says in by the spring of 1954, UNIVAC had
installed 20 computers and IBM only 15.
One correction: the University of Manchester/Ferranti ATLAS computer was not
the first computer to have an interrupt system. Instead it appears to have
been the first computer to have hardware-based virtual memory. Knuth
_suggests_ that interrupt systems were invented by E. W. Dijkstra in 1957,
but this date seems suspiciously date since sometime around 1958 IBM was
marketing a computer (the IBM 709) which had an interrupt system.
P.S. A stray thought:
My brief research on this matter has produced the following tentative
conclusions:
1) Turing and von Neumann were independent co-inventors of the stored-program
computer the way Newton and Leibniz were independent co-inventors of the
calculus.
2) Neither Turing nor von Neumann were interested in a priority dispute.
From the point of view of a math or computer historian, 2) is unfortunate,
because if it were otherwise there would be a good deal of evidence placed in
the public domain!!
James A. Landau
systems engineer
FAA Technical Center (ACT-350/BCI)
Atlantic City Airport NJ 08221 USA