XV
Many researchers, including myself, concur with the idea that one should first
grasp the logical principles (of, say, a software project) before working out the
“engineering” details. But I object when people — and even the community at
large — rephrase history in support of their cause. Robinson, for example, also
writes that “Turing’s 1936 idea had started others thinking” and that “[b]y 1945
there were several people planning to build a universal [Turing] machine.” [99,
p.10]. But, in fact, most people during the 1940s did not know what a universal
Turing machine was, nor had they come across Turing’s 1936 paper or a recast
version of his work.
Alas, also trained historians have fallen into the trap of merely asserting,
rather than proving, Turing’s influence on computer building. Lavington, for
example, writes in his recent book Alan Turing and his Contemporaries: Building
the world’s first computers, that
Turning theory into practice proved tricky, but by 1948 five UK research groups
had begun to build practical stored-program computers. [66, p.xiii, my empha-
sis]
All of the designers of early computers were entering unknown territory. They
were struggling to build practical devices based on a novel abstract principle
— a universal computing machine. It is no wonder that different groups came
up with machines of different shapes and sizes, [. . .] [66, p.8, original emphasis]
These words contradict what ironically seems to be Lavington’s main thesis:
that Turing had almost no direct influence in computer building [66, Ch.8]. A
similar, yet more severe, critique holds for Mahoney’s entire oeuvre — Histories
of Computing [71] — which I have presented elsewhere [43].
As the previous examples illustrate, getting Turing’s legacy right isn’t easy.
By hardly mentioning Turing in their joint book, Computer: A History of the
Information Machine [25], Campbell-Kelly and Aspray put Turing into the right
context. Similar praise holds for Akera’s prize-winning book Calculating a Nat-
ural World: Scientists, Engineers, and Computers During the Rise of U.S. Cold
War Research [8]. (A critical side remark here is that Akera’s book is also about
computing in general and, as the present article shows, Turing’s work did play
an important role in this more general setting.) The research of De Mol [75], Ol-
ley [92], Priestley [94, 95], Daylight [41, 45], and the promising work in progress
by Haigh, Priestley, and Rope all attempt to
1. clarify what we today call the “stored program” concept, and/or
2. put Turing’s and von Neumann’s roles into context by explaining what these
men did do or how their work did influence other computing pioneers.
Finally, although Mahoney uncritically documented Turing’s legacy [43], he
also presented an impressive coverage of the rise of “theoretical computer sci-
ence” — a topic that, to the best of my knowledge, has so far only been scruti-
nized by Mounier-Kuhn [79–82]. Two findings of Mounier-Kuhn that complement
the present article are, in brief, that
1. Turing’s work, and modern logic in general, only gained importance in French
computing during the 1960s, and
XVI
2. it is a “founding myth of theoretical computer science” to believe that the
“Turing machine” was “a decisive source of inspiration for electronic com-
puter designers” [81].
5
Closing Remarks
The existing literature is rather vague about how, why, and when Turing assumed
the mantle of “the father of computer science”. This article has partially filled
that gap by showing that Turing’s 1936 paper became increasingly relevant to
the influential ACM actors Carr and Gorn around 1955. Alan Perlis’s reception of
Turing has yet to be documented in future work.
Carr and Gorn were inspired by re-cast versions of Church, Post, and Turing’s
original writings. These versions included:
– Rosenbloom’s 1950 Elements of Mathematical Logic [100],
– Kleene’s 1952 Introduction to Metamathematics [64], and
– Markov’s 1954 Theory of Algorithms [72].
In the late 1950s, also British automatic programmers, including Booth and Gill,
appropriated ideas from Turing’s 1936 paper.
All aforementioned men were attracted to the “universal Turing machine”
concept because it allowed them to express the fundamental interchangeability
of hardware and language implementations (of almost all computer capabilities).
Unsurprisingly, insights such as these came during a decade of cross-fertilization
between logic, linguistics, and programming technology — a decade in which
computing was still a long way from establishing itself as an academically re-
spectable field.
Just like some space cadets, also the reader may have noticed a strong sim-
ilarity between Weaver’s 1949 “interlanguage” or “universal language” on the
one hand, and the notion of an intermediate machine-independent programming
language on the other hand.
45
Another example of technological convergence is
the pushdown store which, as Oettinger explained in his 1961 paper, served a
unifying technological role across the domains of automatic programming and
machine translation [89].
Understanding the birth of computer science — a term which I use here
for the first time — amounts to grasping and documenting the technological
convergence of the 1950s. On the theoretical side of this convergence, Church’s,
Turing’s, and especially Post’s work became increasingly relevant — a topic that
concerns the notion of undecidability (see [41, Ch.2]) and which lies outside the
scope of the present article. On the practical side, the universal Turing machine
played a clarifying role in that it helped some experienced programmers, the
space cadets, to grasp the bigger picture of what they had been accomplish-
ing in conformance with the language metaphor (— a metaphor that became
well established during the 1950s). To be more precise, the universal Turing
machine paved the way for the complementary onion-skin metaphor, thereby
allowing the space cadets to view their translation problem (from one language