ARTICLE

How Percy Ludgate’s 1909 Paper (and IBM)

Helped Thwart Konrad Zuse’s Computer
Patent in 1960
Brian Coghlan , Trinity College, D02 X9W9, Dublin, Ireland
Brian Randell , Newcastle University, NE4 5TG, Newcastle upon Tyne, U.K.
Ralf Buelow , Heinz Nixdorf MuseumsForum, 33102, Paderborn, Germany

This investigation discusses how Percy Ludgate’s paper from 1909, which details
his design for a mechanical computer called the “Analytical Machine,” was utilized
by a German patent attorney in 1960 to prevent the approval of Konrad Zuse’s
computer patent, just in time, as the patent (first applied for on 16th June 1941) was
about to be granted. This narrative is then followed by analyses of, and discovery
of proof of, the role of IBM in preventing Zuse from getting what would have been
the premier patent on the concept of a programmable computer, and then by an
exploration of how information about Ludgate’s 1909 paper was found by (or for)
the opposing German patent attorney.

T
here were only two mechanical designs pub- that specifically stated inheritance from either of their
lished for a programmable computer well before machines. Hence, Babbage and Ludgate’s importance
the electronic computer era, Charles Babbage’s has been seen as primarily historical. However, in 2022
“Analytical Engine,” details of which were published by one of the authors, Ralf Buelow of the Heinz Nixdorf
Ada Lovelace in 1843 [3], [33], and Percy Ludgate’s “An- MuseumsForum, outlined in [12]1 how Percy Ludgate’s
alytical Machine,” published by Ludgate in 1909 [15], 1909 paper design for a mechanical computer was
[16], [34], [35], [55]. Subsequently, from c.1914 Leonardo used in 1960 by a German patent attorney to thwart
Torres y Quevedo and his successors began electro- Konrad Zuse’s computer patent just as it was about to
mechanical designs, and from c.1937 electronic de- be granted, and how Zuse thereafter alleged that the
signs began [43], then from c.1949 fully electronic mod- attorney had been helped by IBM.
ern computers and their billions of successors started This article describes a detailed investigation we
to emerge. A third mechanical design dates from the have made of these matters. For technical details of
dawn of the electronic computer era in 1936, namely the design of the Z3 machine, and of the claims in
Konrad Zuse’s “V1,” later renamed “Z1” [61]—his sub- Zuse’s patent application, we refer the reader to the
sequent designs, and particularly the Z3 (1941), which excellent accounts by Raúl Rojas [46], [47], [48]. Our
features large in this account, were electromechanical. paper thus is largely based on translations of the rel-
Until recently it appeared that Babbage and evant sections of the principal documentary evidence
Ludgate had no influence on modern computers, since we have found relating to this patent dispute, includ-
as far as was generally known there was no record in the ing that related to IBM’s involvement in the successful
literature concerning modern electronic computers effort to prevent Zuse obtaining what might have be-
come the primary patent of the concept of a program-
mable computer. We may now say that Ludgate, whose
© 2024 The Authors. This work is licensed under a Creative work hitherto had seemed merely of historical interest,
Commons Attribution 4.0 License. For more information, in fact may well have had some influence on the way in
see https://creativecommons.org/licenses/by/4.0/ which the international computer industry developed,
Digital Object Identifier 10.1109/MAHC.2024.3369024
Date of publication 23 February 2024; date of current version
1
12 September 2024. English translation (2023) of [12] available at [56].

20 IEEE Annals of the History of Computing Published by the IEEE Computer Society July-September 2024
 ARTICLE

even though his work was unknown to the designers Only in late 1952 was it made public with fif-
of the first electronic computers. ty-one claims. The examiners had - eleven
years after filing - no problems whatsoever
ZUSE’S PATENT LITIGATION with the patent eligibility. None of the com-
All of the Zuse patent litigation documentation, and to puter manufacturers who are important to-
date almost all of the literature about it, is in the German day filed petitions against the patent, only
language. The microfilm images of the litigation docu- Triumph, which however was later backed
ments are published online by the Konrad Zuse Internet by IBM. With admirable attention to detail,
Archive [30], [31] and scanned images by the Deutsches a whole host of objections were collected
Museum [21]. The most extensive literature about the - clearly a great contribution to the history
litigation is the excellent German-language account of computing. Yet the innovation and pro-
given by Hartmut Petzold’s chapter “Die Mühlen des gressiveness of my invention still could not
Patentamts” (“The Mills of the Patent Office”) [41]2 in be doubted. Only in mid-1967, that is, twen-
Rojas’ very illuminating book “Die Rechenmaschinen ty-six years after filing the patent, was a fi-
von Konrad Zuse“ [45]. For ease of comparison by future nal decision made by the German Patent
scholars, for the principal original German-language Office. It said in effect that a patent-worthy
litigation documents mentioned in this article we have invention did not exist: “The innovation and
prepared a set of online supporting documents in progressiveness of the object concerned in
which the original German texts are shown side-by-side the main application are not doubted. Yet a
(for 2-page view) with very careful translations to En- patent cannot be granted due to insufficient
glish [56]; these translations are the source of the great inventive merit.” Not only I, but also the Tele-
majority of the quotations in the present text. funken Company was deeply disappointed. I
Konrad Zuse applied for many patents on his com- had a patent contract with them, and their
puting inventions.3 It must be stressed that this sec- patent department had worked on my case.
tion concentrates entirely on Zuse’s efforts in relation I was also pained to learn that the deciding
to just one of these applications, his attempt to obtain appellate court had not allowed the appeal
a patent for the concept of a programmable digital to reach the Bundesgerichtshof, Germany’s
computer based on his Z3 computer [60]. It expands Federal Court. [61, pp. 109–110]4
on Petzold’s account, specifically enlarging on those
elements that relate to Percy Ludgate’s 1909 paper on Zuse first registered the principle of the Z3 under the
his Analytical Machine. Zuse himself, in his autobiog- title “Calculating device“ with the Reich Patent Office on
raphy, gave this account of his patenting efforts: 16th June 1941, when it was given the file number Z26476
IXb/42m. Only one drawing of the Z3 survived the Sec-
I had filed a few patent applications before ond World War, the machine itself being destroyed in
and during the war, such as the one in 1936 a bombing raid. Zuse escaped from Berlin, bringing his
for mechanical switching techniques with a only surviving machine, his Z4, with him to Bavaria. Prob-
mechanical store - for which I was later giv- ably due to World War II, the application was not pro-
en a patent, albeit only after these designs cessed to a conclusion. After World War II the German
were no longer of any value. I filed my first Patent Office was reorganized, and hence from Novem-
patents for the program control and arith- ber 1951 Konrad Zuse revived his attempt to obtain this
metic units in 1937. They were rejected due patent, which was renumbered Z391, but to Zuse’s very
to insufficient disclosure. During a simulta- evident frustration this application then followed a diffi-
neous filing in the United States, Babbage cult and tortuous trajectory, as is briefly outlined below.
was cited against me - as already mentioned. The Triumph-Werke company in Nuremberg op-
In mid-1941, when Z3 had taken concrete posed this patent throughout. They produced mo-
form, I filed a patent for it. The patent was torcycles, typewriters, and accounting machines [18],
later recorded under file number Z391, and [38]. The patent attorneys working for Triumph-Werke
it was probably my most important patent.
4
The German edition of Zuse’s biography refers on page 98
line 10 to “Bundespatentgericht.” In the English edition [61],
2
English translation (2023) of [41] available at [56]. on page 109 lines 26-27, this is translated in error to “German
3
Coghlan, B., Buelow, R., Randell, B., “Konrad Zuse’s Computer Patent Office” as above, whereas it should be “Federal Patent
Patents,” available at [56]. Court.”

July-September 2024 IEEE Annals of the History of Computing 21

ARTICLE

collected older patents that Zuse’s claims conflicted numbers, so that there is also no command
with, so causing rejection of these claims.5 Zuse then encoding or command decoding and no se-
amended his patent application and resubmitted it in lection mechanism for the alternate connec-
1958, having reformulated the claims as follows: tion of individual machine parts to each oth-
er. [56, letter of 30th July 1959, p. 219]
In contrast, the invention is characterised by
the combination of the following features, The Patent Office declared that they could not find
some of which are known per se, as men- in the works of Babbage, Louis Couffignal,6 Torres y Que-
tioned above: I. At least one memory unit for vedo or George Stibitz any program-controlled calculat-
storing the numbers in specific memory cells, ing machine whose construction and mode of opera-
II. at least one arithmetic unit, III. a planning tion corresponded to Zuse’s claims, and therefore they
unit for controlling operational sequences in intended to grant the patent (although as can be seen
accordance with a predetermined program, below, later the Patent Office changed its judgement re-
which cooperates with a device suitable garding the French computer pioneer Couffignal’s work):
for the output of character combinations in
such a way that the program, developed in Through the work of Ch. and H.P.Babbage,
accordance with operational and transmis- the idea of the program control of calculating
sion processes, runs successively as an or- machines by means of punched cards, such
dered sequence of individual commands en- as those used for the control of looms, has
coded in the form of numbers or the like. IV. become known in principle, and special cir-
and at least one selecting unit for decoding cuits and devices have also been developed
the set instructions and initiating the asso- by L. Couffignal for the automatic solution of
ciated arithmetic operations and transmis- certain arithmetic operations, but precise
sions by connecting the called memory cells information on the constructive solution of
to setting or result elements of the arith- this problem, as it is to be solved by the pro-
metic unit, such a selecting unit containing gram-controlled calculating machine form-
more active elements on the output side ing the subject of the present application, is
than there are setting elements on the input not contained in any of the literature. Also, in
side. [56, letter of 8th May 1958, p. 430] the solution derived from the work of Torres
y Quevedo (Bulletin de la société d’encour-
Note that item IV above specified encoded instruc- agement pour l’industrie nationale) as well
tions, highlighted in the description as a point of demar- as that from the report by G.R.Stibitz on the
cation from Babbage, which for Zuse became a focus of Bell Complex Number Computer (Bull.Amer-
the subsequent litigation, as will become evident below. ican.Mathem.Soc.) there is no program-con-
In 1959, the examiners of the Munich-based Pat- trolled calculating machine to be found
ent Office voted to approve Konrad Zuse’s application, whose construction and mode of operation
despite the objections that had been made to it, and are identical with the machines forming the
wrote to Zuse’s patent attorney stating: subject-matter of the present application.

The patent specifications cited by the oppo- The patent specifications and literature cit-
nent thus relate to individual parts of the cal- ed therefore do not prevent the present ap-
culating machine which is the subject-mat- plication from being patented. […]
ter of the present application, none of these
patent specifications describe a machine It is therefore intended to grant the re-
which contains all of these parts in such a quested patent on the basis of the docu-
combination as is indicated by the new main ments submitted [56, letter of 30th July 1959,
claim filed on 13 May 1958. Furthermore, in pp. 220–221]
none of the machines and devices described
in the above-mentioned patent specifica- 6
Louis Pierre Couffignal proposed an electromechanical
tions are the commands given in the form of computer in 1938 [43], and later headed the digital comput-
ing laboratory of the Institut Blaise-Pascal. He visited Aiken
and von Neumann in 1946, and on 20 May 1947 visited Zuse
5
[56, letter of 29th November 1957, pp. 422–423]. in Bavaria [7].

22 IEEE Annals of the History of Computing July-September 2024

 ARTICLE

FIGURE 1. Left: page 1 of Hagen’s letter to the Munich Patent Office of 11th January 1960 citing Ludgate. right: Weber–Schäfer’s let-
ter to Zuse of 28th January 1960, referring to Hagen’s letter of 11th January 1960. Courtesy of the Konrad Zuse Internet Archive [30].

The importance of this letter cannot be overstat- analytical machine”, published in the journal
ed, and in January 1960 Konrad Zuse’s computer pat- “Scientific Proceedings Royal Dublin Soci-
ent was about to be granted. However, on the 12th of ety”, 1909, Volume XII, Issue IX, Pages 77-91
January 1960, the Patent Office received a new ob- which had not yet been taken into account
jection by the company Triumph-Werke, sent the day in the previous examination procedure.
before, which constituted somewhat of a bombshell. [56, letter of 11th January 1960, p. 206]
The trajectory of the patent litigation then sharply re-
versed. Triumph-Werke had engaged a new patent at- Hagen also mentioned Charles Babbage [3], [33],
torney, Dr. Gerhard B. Hagen who in the first mention Couffignal [19], and Claude Shannon [49]:
of Ludgate thus far referred to his 1909 paper, giving a
precise reference and a ten page explanation of it, in a The most significant objection in the exam-
letter (Figure 1) to the Patent Office starting: ination procedure so far was the description
of the computing system by Ch. Babbage […]
In reply to the decision of 30.7./11.9.1959. Ludgate’s publication shows that although
he was not familiar with Babbage’s calcu-
The applicant filed new patent claims 1 - 5 lating machine when he planned his device,
by petition of 12.5.1958, which the examining he used the same basic principle […] Such an
office initially considered to be allowable. arithmetic unit was the subject of an earlier
In this opinion, however, these claims are application, the applicant’s Application Z 23
not patentable, since claim 1 is identically 624 IX/42m and is in itself an arithmetic unit
anticipated and claims 2-5 are not inven- similar to the structure of the French patent
tive, both deficiencies being based on the of Couffignal 819 695 […] Couffignal’s arith-
literature of P.E. Ludgate “On a proposed metic unit, which works in binary form, is not

July-September 2024 IEEE Annals of the History of Computing 23

ARTICLE

only capable of multiplication and division, For these reasons, in our view, the appli-
but also of extracting roots […] This arith- cation should be rejected in its entirety.
metic unit plays the role of the mechanism, [56, letter of 11th January 1960, pp. 213–215]
which is called “mill” by Babbage and also by
Ludgate […] Both Babbage and Ludgate used Whether Zuse knew of their work was unimport-
storage mechanisms as an essential part ant; the decisive factor was that someone had the
of their calculating machines […] Likewise, ideas before him. Telefunken’s patent attorney Dr.-
in Babbage’s and Ludgate’s machines, the Ing. Max Weber-Schäfer, acting for Zuse,7 sent him a
arithmetic unit and the memory unit were worried letter (Figure 1) on the 28th of January 1960,
interconnected by a program unit […] The stating that he believed that “a whole new situation
use of selection pyramids is shown in the has been created,“ as it was “an objection to be taken
aforementioned French patent specification seriously“:
of Couffignal […] they are also dealt with in
the paper of Shannon. [56, letter of 11th Jan- Betr.: Z 391 IX/42m (FA.W. 541)
uary 1960, pp. 208, 209, 211, 212, 214]
Dear Dr. Z u s e !
But Hagen principally focussed on Ludgate’s paper,
ending after ten pages of closely worded objections In this matter, the petition from Triumph has
with: now been received, which I have already
reported to you orally that the examiner
According to this view, the new claim 2 is had promised. Through this input it seems
also completely anticipated by the Ludgate to us that a whole new situation has been
publication. created, because of a new reference to
P.E. Ludgate “On a proposed analytical ma-
Claim 3 emphasises the use of binary code for chine,” from Scientific Proceedings Royal
the layout; here, too, no protectable disclo- Dublin Society, 1909, Volume XII, Issue IX,
sure can be seen, since coded punched tape pages 77-91, of which we enclose a pho-
control is based precisely on the derivation of tocopy for you. We have not yet examined
the control processes from yes-no circuits. the reference in detail, but after a cursory
glance at the sources named in the brief, it
Claim 4 says no more than what is shown in seems to us that this is an objection to be
Swiss patent specification 181 926 with re- taken seriously.
spect to the punched tape control, applied
to a telegraph system. There, too, selection We would like to ask you to study the ob-
of circuits and selection of letters and nu- jection as soon as possible and let us know
merical values take place. what you think about the possibilities of lim-
itation. For our part, we will also take a close
The new claim 5 refers to the use of selec- look at the literature and consider it expe-
tion pyramids. The use of selection pyramids dient for us to discuss the matter as soon
is shown in the aforementioned French pat- as possible after your return from your trip.
ent specification of Couffignal […] they are The deadline for replying to the pleading is
also dealt with in the paper of Shannon, pub- 22.3.1960. On our part, having asked the
lished in the journal “Transaction of Ameri- examiner to speed things up, we would like
can Institute of Electrical Engineers”, 1938, to avoid exceeding the deadline or asking
page 713, volume 57. for an extension. [56, letter of 28th January
1960, p. 439]
With respect to the other claims still main-
tained, […] since the subclaims cannot form Because Weber-Schäfer had attached a copy
a further development of the idea character- of Ludgate’s paper, Zuse was able to study it and
ized in the current claim 1, they cannot serve draft a rebuttal (Figure 2). Zuse concentrated on
as a basis for a possible restriction of claim
1, which is not to be regarded as eligible for 7
Telefunken supported Zuse’s attempt to patent the princi-
protection. ples of his Z3 computer.

24 IEEE Annals of the History of Computing July-September 2024

 ARTICLE

FIGURE 2. Left: German original of page 2 of Zuse’s response of 2nd March 1960 to Hagen’s letter of 11th January 1960. Right:
German original of page 1 of Hagen’s letter to the Federal Patent Court of 17th November 1966. Courtesy of the Konrad Zuse
Internet Archive [30].

the perforated paper, which carried the program of On 29th March 1960, Weber-Schäfer drafted a
Ludgate’s machine. Zuse saw no counterpart to this new patent application8 aimed at countering Hagen’s
in his patent application. He categorically ruled out a now-revised objections, and in hindsight this might be
“Vercodungssystem” (coding system) as being prior art considered the beginning of the end.
by Ludgate, writing: In 1961, Hagen filed a further objection against
Zuse’s now-revised patent application, based on Louis
It corresponds to a wrong interpretation of Couffignal’s 1938 doctoral thesis [19]. Triumph-Werke’s
the writing of LUDGATE and is the product objections were successful. In a preliminary letter
of the imagination of the opposing party. [56,
from the Patent Office to Telefunken patent attorney
letter of 2nd March 1960, p. 442]
Dr.-Ing. Benno Johannesson (acting for Zuse) on 15th
November 1961 and a confirmation on 20th September
Zuse may have made an error, as Ludgate’s 1909
1962, the Munich Patent Office issued a final refusal to
paper does not specify the format of instructions, only
patent Zuse’s concept.
stating:

Thus, claim 1 cannot be granted either in the

Each row of perforations across the formu-
la-paper directs the machine in some defi- version of 28 April 1960, or in an amended
nite step in the process of calculation—such version which is limited to a device for trans-
as, for instance, a complete multiplication, ferring numbers from the arithmetic unit to
including the selection of the numbers to the memory and vice versa.
be multiplied together. [35, p. 80], 28th April
1909 8
[56, letter of 29th March 1960, pp. 443–449].

July-September 2024 IEEE Annals of the History of Computing 25

ARTICLE

In this situation, the requested patent is like- arithmetic unit and a program unit which
ly to be refused. [56, letter of 15th November executes the operation and transport com-
1961, p. 169] mands on a program carrier. According to
the second paragraph of the new introduc-
After proper publication and examination of tion to the description of 22 June 1964, it is
the objection, the request for this patent to based on the task of developing measures
be granted “Program-controlled calculating for the construction of such machines,
machine” failed. [56, letter of 20th Septem- which have only been described so far, but
ber 1962, p. 152] not implemented.

In reaction, in 1962 Konrad Zuse took the case one 2. A task which is itself inventive is obvious-
step further by having his attorney appeal to the then ly not present. […] Consistent with this, the
newly established Federal Patent Court, which was applicant, in defending the subject-matter of
also based in Munich, this time more specifically: the application, did not rely on the status of
the task, but on the means identified in the
This construction principle, […], is based on claim, with which it intends to solve the task.
the idea underlying the invention of using
an instruction code encoded in yes-no value 3. According to claim 1 of the main applica-
combinations in program-controlled calcu- tion, in accordance with the applicant’s sub-
lating machines. missions, general protection is sought for
the concept of the programmable feature of
In none of the previous publications relating the machine described in the preamble, con-
to program-controlled calculating machines sisting of controllable switch groups which
- these are the works of Babbage, Ludgate contain more output control lines than input
and Couffignal - is this idea explicitly stated. control lines, i.e., that is constructed accord-
[56, letter of 19th December 1962, p. 140] ing to a so-called Christmas-tree circuit (e.g.,
German patent specification 554 888). […]
Subsequently, Zuse eventually reformulated the
patent specification, stating in its introduction: The novelty and progressiveness of the sub-
ject-matter claimed in the main application
However, nothing more can be learned from cannot be doubted. However, for lack of
this proposal about the technical means for inventive value, no patent can be granted.
program control than from the older works [56, letter of 14th August 1967, p. 5, pp. 7–9]
of Babbage and Ludgate. [56, letter of 22nd
June 1964, p. 98] With this acceptance of the relevance of the prior
work by Couffignal, Babbage, and Ludgate, and the
The Federal Patent Court, after considering Zuse’s expressive phrase, “for lack of inventive value, no pat-
complaint, rejected his appeal on the 14th August ent can be granted“ the Patent Court ended the Zuse
1967. Their judgement stated: patent proceedings, in which Zuse’s computer patent
application was refused twice, in 1962 by the Patent
The general concept of claim 1 corresponds Office and in 1967 by the Patent Court, and in which
to the literature cited on program-con- Percy Ludgate’s 1909 paper played a crucial role at the
trolled calculating machines by Couffignal most critical time, just as the patent was about to be
(Dissertation 1938), Babbage and Ludgate in granted in January 1960.
the objection. […] Zuse subsequently wrote some pithy comments
on German patent law:
The appeal is admissible, but could not
succeed. It must be considered a failure of our patent
law that a patent application can take so long
1. In the preamble of the patent claims, the that the general application of the inventive
application refers to program-controlled idea, at the time quite amazing, has since
calculating machines with a memory, an become quite usual. […] In the early days of

26 IEEE Annals of the History of Computing July-September 2024

 ARTICLE

computer development, on the other hand,

ZUSE AND IBM
there was great confusion about patent pro-
Representatives of IBM’s subsidiary in Germany, De-
tection for computers and all that went with
homag,11 first approached Zuse when he lived in the
them. I can indeed count myself – without
late 1940s in Bavaria, visiting him in 1947 [37, p. 37]. In a
wishing to be melodramatic about it – among
passage from his autobiography [61] he explains:
the victims of this confusion. [61, pp. 111–112]

In Hinterstein I had given [Helmut O. Goeze]

One can but speculate as to what would have been
a demonstration of the Z4, and it made a
the consequences, on Zuse’s career, and indeed the
powerful impression on him. […] He suc-
global computer industry, should the German Patent
ceeded in attracting the attention of the
Court have ruled in Zuse’s favor and granted his patent.
then President of IBM, Thomas Watson, in
The concluding paragraph of Hartmut Petzold’s “The
me and my Z4. Watson then passed on the
Mills of the Patent Office” simply states:
instructions to Hollerith Germany. […] Two of
Hollerith’s managers, Hummel and Scharr,
It is hardly possible to answer the question
headed off in one of the few company cars.
of what the granting of the patent in 1967
[…] Negotiations continued, and finally an
would have meant. At that time, not only
options contract for the acquisition of my
the opposing party had recognised the “out-
patents was concluded. […] Much to our
standing achievement of being the first to
chagrin, IBM was interested solely in my
realise a programmatic calculating machine
patent rights. In the eyes of the Americans,
in practice”; in the end, however, it was sole-
that we wished to continue work on the Z4,
ly a matter of the legally recognized “inven-
even wanted to undertake new develop-
tive strength.” The financial consequences
ments, were “perfectly foolish conditions.”
depended on its recognition. The assump-
[…] Clearly I could have made a good deal on
tion that the entire international computer
my patent rights; and then IBM would have
industry had to pay Zuse license fees does
been on my side in subsequent disputes. […]
not seem very realistic. It must also remain
In the face of such incalculability, I finally de-
open whether the granting of the patent
cided to continue to build up my company.
would have given Zuse KG relief in its diffi-
[…] I probably had no other choice. IBM would
cult financial situation and prolonged its
not even guarantee that I would be able to
period of independence. In 1994, Zuse char-
continue to work in the computer field.
acterised the rejected patent succinctly: “All
[61, pp. 114–115]
computers would have been covered under
this fallen patent.” He said nothing about the
Zuse withdrew from these negotiations, choosing
consequences expected at the time from a
instead to develop his company. Then in about 1949
granting of the patent. [41, p. 107]9
[37, p. 39], he contracted with Remington Rand12 to
build a punched-card machine, the M9 (Mithra-9).13
Zuse might be considered quite unlucky to have
Konrad Zuse believed for the rest of his life that
started so early in 1936 yet failed 31 years later in 1967.
Triumph-Werke had received help from IBM [61, p. 109
On the other side of the coin, Ludgate and Couffignal
(and, as detailed below, IBM) may also now be consid-
ered a European counterpart to the American Atana- 11
In April 1949 IBM’s German subsidiary, Dehomag, an ab-
soff and Honeywell in their (in hindsight worthy) roles breviation of Deutsche Hollerith-Maschinen GmbH (i.e.,
in thwarting or invalidating fundamental European and German Hollerith Machines LLC), changed its name to IBM
American patents, which might have led to the costs Deutschland.
12
The c.1949 contract was for Zuse to provide “calculating
of computers being significantly inflated.10 punches” to Remington Rand, Norwalk, Connecticut [61, p.
115]. There is a curiously triangular involvement of Zuse &
IBM, Zuse & Remington (later Sperry Rand), and IBM & Sperry
Rand.
9 13
English translation (2023) of [41] available at [56]. In the biography of Dudley Buck (MIT Whirlwind team mem-
10
Sperry Rand have been reported as asking for royalties of ber, inventor of the Cryotron and of content-addressable
$250 million from Honeywell at first, reducing to $20 million memory) [22, p. 58] it is suggested the c.1949 contract with
before litigation began, and a total of $150 million from other Remington Rand was part of “Operation Paperclip” c.1949-54,
computer manufacturers [50], [57]. Other reports state 1.5% and was aimed at attracting Zuse to America [11], [37, p. 50],
royalties [32]. [39], [61, p. 116], and so illustrating their high regard for him.

July-September 2024 IEEE Annals of the History of Computing 27

ARTICLE

and pp. 114–115], something that is clearly stated in this under the direction of James Bryce16 (who later knew of
excerpt from Uta Merzbach’s interview of Zuse in 1968: Babbage’s work [1]), as described by IBM’s own history
webpages:
UM: Who represented the opposition. Who
was their lawyer? By hiring engineering consultant [James
Wares] Bryce in 1917, Watson showed that
KZ: That was, I believe, Hagen. Hagen was he recognized the importance of pure in-
his name. venting. Rather than developing products,
Bryce’s job was to dream up new ways of do-
UM: Who did he represent? ing things and patent them. He established
a patent development department in 1932,
KZ: Yes, officially Triumph Corporation, but hiring [Arthur Halsey] Dickinson17
unofficially, IBM. He was with IBM. [37, p. 31]
There are very few mentions of IBM in the volumi-
After the above rejection of IBM’s overtures in 1947, nous Zuse patent litigation records. In fact, the first
IBM would not have wanted Zuse’s patent, or probably explicit mention of IBM in the Zuse patent litigation
any computer patent that might place IBM at a competi- appears to be in 1963 in a letter to Zuse from his patent
tive disadvantage, to be granted.14 In 1953–1954, IBM con- attorney Weber-Schäfer:
sidered contesting the ENIAC patent by claiming Atana-
soff and Berry’s ABC technology as prior art [50]. But in Relating to these two applications, according
1956 Sperry Rand, who by then owned the ENIAC patent to the attachments, IBM has asked for per-
rights, cross-licensed patents with IBM.15 In 1959 IBM mission to inspect the files. We will not ob-
began research and development on their 360 series, in- ject to their request. [56, letter of 16th August
troduced in 1964, the same year that ENIAC’s American 1963, p. 476]
patent [24] was granted. The conundrum for IBM would
have been how to stop Zuse’s patent using evidence of The next mentions began two years later when
prior work (e.g., Ludgate 1909) without that evidence Zuse’s patent attorney Johannesson attempted to
compromising the viability of the ENIAC patent. amplify statements made in an essay by a senior IBM
During this period, Hagen’s correspondence with researcher, that:
the Patent Office over this patent application gave no
indication of any involvement of IBM. However, Hagen the applicant would like to refer to the sub-
was in fact acting for IBM Deutschland at this time in sequently published article on the question
other patent matters. For example, even during the of the technical progress achieved by the
critical period 1959–1960, Hagen simultaneously acted invention on which this application is based:
for an IBM patent application filed (given as for IBM Dr. Karl Ganzhorn “Historical Development of
New York) on 16th December 1959 [27] and another Information Processing” in Issue 164 of IBM
filed on 14th September 1960 [28]. News, Volume 14, February 1964, pages 2152
It is evident that IBM took patenting very serious- to 2156 [56, letter of 22nd October 1965, p. 66]
ly, being very active in patenting from 1932 onwards
Dr. Ganzhorn’s essay [25], which favorably men-
tioned Zuse and his Z3, was in IBM Nachrichten, the
14
For example, Hasler AG (now Ascom) planned to market public-relations magazine of IBM Deutschland. Gan-
ERMETH (an electronic replacement for the Z4) worldwide
[10] under a 1954 license from ETH (Eidgenössische Tech- zhorn was born in 1921 in Sindelfingen near Stuttgart
nische Hochschule, Zürich), potentially with a sublicense and died there in 2014. He was a physicist, and was
from Zuse. At the end of 1955, before ERMETH’s completion, founder and long-time Director of the IBM Deutsch-
its principal designer Ambros Speiser [51], one of the two
postgraduates who ran the Z4 rented by Professor Eduard land laboratory in Böblingen.
Stiefel from Zuse [9], left ETH to establish and lead IBM’s
Zürich Laboratory in nearby Rüschlikon, so thwarting Hasler
16
AG’s plans, leading to discord and total loss of interest in ER- Bryce’s US patent specification 2 141 598 was cited in a let-
METH [8]. ter by Hagen [56, letter of 2nd December 1965, p.64] and in a
15
The cross-licensing agreement between Sperry Rand and dismissive response by Weber-Schäfer [56, letter of 20th De-
IBM was signed on 21 August 1956 [57]. IBM has been re- cember 1965, p. 58].
17
ported as paying royalties of $10M [32], $1M/year [40], or apollotv.online, also accgian’s blog and equant.org, archived
$1M/year over 8 years [52], with $200k of $10M relating to extracts of retired IBM webpage “IBM Patents and Innova-
ENIAC [44]. tion,” available at [56].

28 IEEE Annals of the History of Computing July-September 2024

 ARTICLE

Following this, in a 2nd December 1965 letter18 to extensively investigated how Ludgate’s 1909 paper was
the Patent Court, Hagen contradicted the praise of disseminated, what evidence exists of it having been
Zuse by Ganzhorn, suggesting Valtat,19 Couffignal, and read and appreciated, and who might have been involved
Bryce as forerunners to Zuse, prompting, on the 7th in passing on information about the paper and his ideas.
January 1966, Johannesson to respond to the Patent Here we summarize our findings and provide a graph-
Court on behalf of Zuse, for the first time specifically ical portrayal of the numerous possible information flow
expressing their perception of IBM’s role in supporting routes we have identified, and of our very subjective as-
the rejection of the patent application: sessment of their likely relative probability. Full details of
this investigation are provided in two online reports, one
The objecting party is in a difficult position, an analysis21 underpinning this summary, the other a fo-
as it obviously receives information to fight rensic investigation of sources,22 in the hope that other
this application from IBM, but cannot deny researchers will be motivated and able to augment our
that an important IBM voice (Dr. Ganzhorn) findings, and indeed come to more definite conclusions
has recognized openly and sincerely the fa- regarding the route from Ludgate to Hagen.
mous ‘Z 3’ as a milestone in computing histo- Ludgate’s plans for an Analytical Machine did at-
ry. [56, letter of 7th January 1966, p. 53] tract some attention in the scientific literature when
they were published, specifically in a short paper “A
In this letter, and a subsequent letter of 3rd October New Analytical Engine” by Professor C.V. Boys [5] in
1966 to the Patent Court20 in regard to an earlier letter the widely circulated magazine Nature commenting
from Dr. Hagen, Johannesson mentioned “Lizenzver- enthusiastically on Ludgate’s ideas, and in the infor-
handlungen (license talks)” between Telefunken and mative entry “Analytical Machine” in the influential
IBM and accusations from IBM that the Telefunken pat- IEE Science Abstracts: Section A.—Physics [29].23 Two
ent department had slowed down the litigation process. popular engineering magazines (Engineering and En-
While this seems somewhat tangential to the views ex- glish Mechanic and World of Science), published short
pressed in the letter above, it may have precipitated an but informative articles based on the 1909 paper with-
explicit disclosure of IBM’s role, since finally, one month in months of its publication, two of which included
later in November 1966, Hagen wrote to the Patent Court an explanatory drawing, almost certainly provided by
(Figure 2), defending the right of companies to seek sup- Ludgate himself. Presumably these various papers and
port from other companies, and for the first time provid- articles, all of which gave full bibliographical details of
ing clear evidence implying that IBM had been helping Ludgate’s paper, helped the initial spread of knowledge
Triumph-Werke, just as Zuse had always alleged. about his machine, but we have found no further actu-
al evidence attesting to this happening.
In the present case, the applicant, Zuse KG, We have established that the Royal Dublin Society
is being advised by Telefunken AG and its em- journal in which the 1909 paper appeared was routine-
ployees. If the IBM company assists the oppos- ly distributed to several hundred other scientific soci-
ing party in an advisory capacity, then the op- eties worldwide, at least a few of which included brief
posing party is simply exercising its right to do listings of the contents of received journals in their
so. [56, letter of 17th November 1966, pp. 14-15] own publications. Ludgate himself briefly mentioned
his work on an Analytical Machine in the chapter on
Babbage’s Analytical Engine that he contributed to the
FROM LUDGATE’S 1909 PAPER TO 1914 Napier Tercentenary Celebration Handbook [34].
HAGEN’S 1960 LETTER All the above information sources were available
Given the low-profile Ludgate’s work had until Randell’s in a number of German libraries (which we detail in our
1971 paper [42] and 1973 book [43], the question arises:
how did Triumph-Werke and its patent attorney learn 21
Coghlan, B., Randell, B., Buelow, R., “How Hagen found
of the existence and obtain a copy of Ludgate’s 1909 Ludgate’s 1909 paper,” available at [56].
22
paper? In an attempt to answer this question, we have Coghlan, B., Buelow, R., Randell, B., “Pre-1971 Known Refer-
ences to and Possible Dissemination of Ludgate’s 1909 pa-
per,” available at [55].
23
Coghlan, B., Randell, B., Buelow, R., “The Patent Attorneys
18
[56, letter of 2nd December 1965, pp. 62–65]. of the Zuse Z3 Patent Litigation,” available at [56], gives some
19
Raymond Louis André Valtat applied for a patent on a binary biographical information on the patent attorneys of the Zuse
calculator in 1931 and published a theoretical analysis in Z3 patent litigation. Two of the principal attorneys, Hagen and
1936 [43]. Hoffmann, had physics degrees and so would be likely to be
20
[56, letter of 3rd October 1966, pp. 18–20]. interested in physics abstracts.

July-September 2024 IEEE Annals of the History of Computing 29

ARTICLE

report), so our impression is that Ludgate’s paper could as can be seen from the extensive bibliography in his
have been relatively easily found in 1960, even though well 1962 book [26]. There is in fact good reason to assume
before the Internet era, if it had been explicitly searched that Hoffman worked closely with Hagen, and was both
for. The questions that remain include who, including equipped and motivated to find and pass on informa-
the scientifically trained Hagen, might have undertak- tion which would help a challenge to the Zuse patent.
en searches, with what motivation, from what starting Thus any of these three individuals may have
points, and when were these undertaken—and also what played a role, not necessarily knowingly, in allowing
subsequent information transmission was involved. Hagen to find out about Ludgate’s work and acquire a
The first search we have learned of was that which re- copy of his article.
sulted in the 1938 Moore School of Electrical Engineering A quite different set of possible routes by means of
Report by Irven Travis “Bibliography of literature on cal- which information about Ludgate’s 1909 paper might
culating machines.” This contains, in its relatively small have reached Hagen involves Ludgate’s 1914 contri-
Arithmetical Machines section, a full bibliographical ci- bution to the Napier Tercentenary Handbook, since
tation for Ludgate’s 1909 paper [53], but no details of the this included a mention of his work on an analytical
paper’s contents. Travis24 later testified that his bibliogra- machine, and a reference to his 1909 paper on it. The
phy was the result of extensive literature searches by stu- Handbook, and Baxandall’s “Calculating Machines
dents under his guidance and had been circulated widely and Instruments: Catalogue of the Collection in the
in 1938 [54]. But it would appear that Travis made no men- Science Museum” (1926) [4] were for many years the
tion of his bibliography, leave alone of Ludgate, when he most prominent and accessible English-language
gave a lecture on “The History of Computing Devices” in sources of information about early calculators and
1946 at the famed Moore School Lectures on “Theory and computation.27 Our investigation turned up evidence
Techniques for Design of Electronic Digital Computers.” indicating the possible separate involvement of either
Another individual who became very knowledge- Leslie Comrie,28 Vannevar Bush,29 Howard Aiken,30
able about the early history of digital computers was or Ambros Speiser14 in passing on information about
Maurice Wilkes.25 His 1956 book [58] contains, in an ex- Ludgate that had been gained via one or other of these
cellent survey and detailed discussion of the origins of two sources.
computers, the only other full citation of Ludgate’s 1909 Some of these routes are distinctly more probable
paper we have found that was published in the period than others. For example, there is an intriguing possi-
1915–1960. Wilkes attended the Moore School Lectures, bility, involving Aiken, Speiser, Hoffmann then Hagen.
his book was a prominent early computer textbook, and Howard Aiken cited Ludgate indirectly in 1937, lectured
our full analyses document a number of interactions he at the Moore School Lectures in 1946, Aiken & Hopper
had which could have provided opportunities for the cited Ludgate more concretely in 1946, then Speiser
transmission of information about Ludgate. spent the year 1949 with Aiken, before returning to
The third individual who is known to have searched ETH Zürich, which rented Zuse’s Z4 from 1949 for five
diligently during this period for information relating to years (for details of all these events see our online
the origins of digital computers was Walter Hoffman,26 reports21,22), after which Aiken visited Zürich c.1950

24 27
Irven Travis, Professor of Electrical Engineering at the The most important other textbooks on the history of calcu-
Moore School of Engineering, Pennsylvania, in the 1930s built lating machines from this period, neither of which was in En-
two differential analyzers, one for the US Army, the other for glish, were Die Rechenmaschine (1925) by Ernst Martin [36],
the Moore School, creating his bibliography in 1938, and later and Le Calcul Simplifié (1928) by Maurice d’Ocagne [23].
28
leading Burroughs Corporation computing. In the U.K., one of Manchester University’s two copies of
25
Maurice Wilkes F.R.S., Director of the Computer Laboratory, Travis’ Bibliography is marked as originating in the Scientific
Cambridge, U.K., in the late 1930s ran its Meccano differential Computing Service, a company set up by Leslie Comrie. Com-
analyzer, but after attending the 1946 Moore School Lectures rie was an avid collector and sharer of literature on compu-
(arriving late and so missing Travis’ lecture) he initiated and tation, was familiar with the Handbook of the Napier Tercen-
led the development of EDSAC. tenary Celebration, and lent literature to Wilkes before the
26
Walter Hoffman became the Head of IBM’s Patent Opera- latter attended the Moore School Lectures. A visit to Wilkes
tions in Rüschlikon, Zürich. From [2]: “Walter Hoffman [was] in Cambridge in 1947 inspired Speiser before he spent 1949
at this time an assistant of Professor Walther in Darmstadt. with Aiken.
29
Walter Hoffman engaged in literature search and literature Vannevar Bush was a renowned Professor at MIT who in the
collection. Later he became a patent lawyer and worked for 1930s created the first differential analyzer, a mechanical an-
many, many years with IBM in one of the patent departments. alog computer for solving differential equations [59].
30
[…] He has a wealth of historical information on developments Howard H. Aiken was Professor of Applied Mathematics at
of this type and also how they eventually led to the state of Harvard University, and conceived the Automatic Sequence
the art which we have now. He helped quite a bit in getting Controlled Calculator (ASCC, funded and built by IBM, first
literature and in getting literature spread around.” operational in 1944, later renamed the Harvard Mark I) [17].

30 IEEE Annals of the History of Computing July-September 2024

 ARTICLE

FIGURE 3. Graph of the more likely possible routes via which information in Ludgate’s 1909 paper might have reached Hagen in
1960.

(meeting Zuse there) [37, p. 50]. In late 1955, Speiser an actual Ludgate patent application, may have ex-
left ETH to establish IBM’s Zürich Laboratory, joined isted somewhere, perhaps not even in the records in
in c.1957 by Hoffmann. The possibility that Speiser in the U.K. or Ireland, and that may have been amenable
1959 alerted Hoffmann and thence Hagen to Ludgate’s to one of the discovery methods available to patent
work, when Zuse’s patent was about to be granted, attorneys (as well as providing much more detail on
also cannot be discounted. Ludgate’s machine). However, we have not attempted
There are however other quite different means by any investigation of this to us rather arcane literature.
which Hagen could have learned about Ludgate. For A graphical representation of the more likely possi-
example, perhaps interested IBM employees knew ble routes by which information about Ludgate’s Ana-
about Ludgate. IBM was well established in Germany, lytical Machine might have reached Hagen, and of their
and since 1956 had an office in Dublin so obtaining differing likelihood, is provided in Figure 3. It shows, for
Ludgate’s paper would not have been difficult; yet John example, that in our judgement the most likely routes are
Moriarty, IBM Ireland’s first graduate engineer and later those that begin with searches that revealed Ludgate’s
the Director of the Computer Laboratory (computing original paper, probably prompted by one or other of the
centre) at Trinity College Dublin, has no memory of any 1909 reviews, and perhaps Wilkes’ 1956 book.
such request and never heard of Ludgate while work- The above analysis at minimum begins the attempt
ing for IBM [14]. Or, although seemingly improbable, to unravel the mystery. There is currently no direct evi-
a Ludgate citation (in another patent application), or dence available on how Triumph-Werke and its various

July-September 2024 IEEE Annals of the History of Computing 31

ARTICLE

patent attorneys actually found and obtained a copy of found out about Ludgate’s Analytical Machine. We
Ludgate’s paper describing his Analytical Machine, let have identified the very few concrete primary referenc-
alone of any information route all the way from Ludgate es to Ludgate in the interval from 1909 to 1960, and po-
(1909) to Hagen (1960), so our analysis remains conjec- tential secondary sources of dissemination, and have
tural. Furthermore, the patent attorneys and IBM might confirmed the previous assumption that prior to 1960,
have had some entirely different route via which they Ludgate’s 1909 paper had a very low profile. This as-
obtained Ludgate’s 1909 paper—we may never know! sumption appeared to remain valid until Randell’s 1971
paper [42]. However, it is now clear that Ludgate’s paper
CONCLUDING REMARKS gained a higher profile within the Zuse patent litigation
Although the potential impact of Zuse’s failure to post-1960.31 The previous assumption also overlooked
gain a patent was previously discussed by Petzold, the wide original distribution of Ludgate’s paper, a dis-
this article uncovers the part played by Ludgate’s pri- tribution which makes the paper’s subsequent obscu-
or publication on his Analytical Machine, and by IBM, rity rather surprising. However, in summary, although
in the fateful patent refusal. It provides much more there is as yet no direct evidence of how Ludgate’s
detail on the players and activities involved in this paper was found prior to its use in 1960, our analysis,21
refusal and broadens the dissemination of the whole drawing on our separate online report,22 represents the
issue from German to English language audiences. beginnings of an attempt to unravel this mystery.
The very fact that Ludgate’s 1909 paper on his Please note that copies of many of the published
Analytical Machine was employed to thwart Zuse’s primary documents are available online at the SCSS
attempt to patent the computer refutes the preva- Ludgate Webpage [55], while supporting documents
lent assumption that Ludgate had no influence on and translations are available via the Index of SCSS
modern computing, and that his work is just of his- Ludgate-Zuse Files [56].
torical interest. Whatever the scale of their influence,
patents can impact the commercial activities in their ACKNOWLEDGMENTS
respective sectors. Hence Ludgate’s paper clearly did For ease of comparison by future scholars, the authors
have a potential influence on the commercial future have provided supporting online documents showing
of computers, even if his work was unknown to the full transcripts of the principal original microfilmed
developers of the early electronic computers. How- German-language litigation documents that have
ever, we have deliberately avoided speculating on the been cited in the paper, side-by-side with very care-
types or the extent of the possible commercial con- ful translations to English by the present authors.32
sequences, either to Zuse’s company, or to the rest of
Figures in this article show images of the actual orig-
the computer industry, of the thwarting of Zuse’s at-
inal pages that we deem to be the most interesting
tempts to obtain what would have been the first pat-
among these documents. The microfilm images of the
ent on a programmable computer. Similarly, we have
original German documents referred to in this article
chosen not to attempt to add to the now extensive
are derived from work by the Konrad Zuse Internet
literature on the technical merits and significance
Archive [30], licensed under a Creative Commons At-
of Zuse’s designs, as compared to subsequent ma-
tribution-NonCommercial-ShareAlike 3.0 Unported
chines such as ENIAC [6], [13].
License [20], for which we offer thanks. The authors
While uncovering this story of how information
also thank Charles Mollan for information on journal
about Ludgate’s Analytical Machine was used, we have
exchanges and accessions in the Royal Dublin Society,
found clear evidence of the actual role played by IBM
the staffs of the ETH Library Zürich, the Bayerische
in the Zuse patent litigation, an activity which Zuse
Staatsbibliothek, and the German Patent and Trade
had only been able to allege, which was something he
Mark Office Library for information on accessions in
did both at the time and thereafter, in particular in his
those libraries, and Deutsche Nationalbibliothek for
autobiography [61]. Indeed, we have shown that IBM’s
role was finally admitted to the German Federal Pat-
31
ent Court in 1966 and have revealed evidence that IBM There are two obvious reasons for this not becoming more
widely disseminated: first that it was confined to patent liti-
had in fact supported the denial of Zuse’s German pat- gation circles, second that within those small circles it was
ent, while having acquired a license to the American further confined to German language speakers.
32
ENIAC patent [52]. One of the authors is a native German speaker, while the
others are native English speakers. The authors do not guar-
We have identified a number of ways in which Tri- antee the correctness of any of these translations, which
umph-Werke and IBM’s patent attorneys might have without prejudice have been done on a best-effort basis.

32 IEEE Annals of the History of Computing July-September 2024

 ARTICLE

hosting their excellent ZDB-Katalog. Finally, the au- [11] R. Buelow, “Konrad Zuse und die Operation
thors extend their thanks for the support of the School Paperclip,” Heinz Nixdorf MuseumsForum, Nov.
of Computer Science and Statistics, Trinity College 5, 2015. [Online]. Available: https://blog.hnf.de/
Dublin, for their support for this work and for the John konrad-zuse-und-die-operation-paperclip/
Gabriel Byrne Computer Science Collection. [12] R. Buelow, “Percy Ludgate und Konrad Zuse,”
Heinz Nixdorf MuseumsForum, Oct. 14,
BIBLIOGRAPHY 2022. [Online]. Available: https://blog.hnf.de/

[1] W. Aspray, “Oral-history: John McPherson,” percy-ludgate-und-konrad-zuse/

Engineering and Technology History Wiki, IEEE [13] A. W. Burks and A. R. Burks, “The ENIAC: The first
Center for the History of Electrical Engineering, general purpose computer (and the subsequent
1992. [Online]. Available: https://ethw.org/ ‘Commentary’, on pp. 389–399),” IEEE Ann. Hist.
Oral-History:John_McPherson Comput., vol. 3, no. 4, pp. 310–389, Oct. 1981.
[2] W. Aspray, “Oral-history: Walter Proebster,” Engineering [14] B. Coghlan, personal email communication with John
and Technology History Wiki, IEEE Center for the Moriarty, 19:22 pm, Oct. 25, 2022.
History of Electrical Engineering, Munich, Germany, [15] B. Coghlan, B. Randell, P. Hockie, T. Gonzalez, D.
Jul. 1993. [Online]. Available: https://ethw.org/ McQuillan, and R. O’Regan, “Investigating the work and
Oral-History:Walter_Proebster life of Percy Ludgate,” IEEE Ann. Hist. Comput., vol. 43,
[3] H. P. Babbage, “The Analytical Engine,” Proc. British no. 1, pp. 19–37, Jan.–Mar. 2021.
Association, (paper read at Bath, Sep. 12, 1888), 1888. [16] B. Coghlan and B. Randell, Percy Ludgate (1883–1922):
[Online]. Available: https://www.fourmilab.ch/babbage/ Ireland’s First Computer Designer. Dublin, Ireland: The
hpb.html John Gabriel Byrne Computer Science Collection,
[4] D. Baxandall, Calculating Machines and Instruments: Trinity College Dublin, 2022.
Catalogue of the Collection in the Science Museum. [17] I. B. Cohen, Howard Aiken: Portrait of a Computer
London, U.K.: Science Museum, 1926. Pioneer. Cambridge, MA, USA: MIT Press, 1999.
[5] C. V. Boys, “A new Analytical Engine,” Nature, vol. 81, [18] Company-Histories.com, “TA Triumph-Adler AG.” 2024.
no. 2070, pp. 14–15, Jul. 1909. [Online]. Available: https://www.company-histories.
[6] A. Brennecke, “A classification scheme for program com/TA-TriumphAdler-AG-Company-History.html
controlled calculators,” in The First Computers— [19] L. Couffignal, Sur l’analyse Mécanique. Application aux
History and Architectures, R. Rojas and H. Hashagen, Machines à Calculer et aux Calculs de la Méchanique
Eds., Cambridge, MA, USA: MIT Press, 2000, pp. 53–68. Céleste. Paris, France: Gauthier-Villars, 1938.
[7] H. Bruderer, “Konrad Zuse’s guestbook: A [20] Creative Commons Attribution-NonCommercial-
treasure trove,” BLOG@CACM, Commun. ShareAlike 3.0 Unported License, 2024. [Online].
ACM, Dec. 2020. [Online]. Available: https:// Available: https://creativecommons.org/licenses/
cacm.acm.org/blogs/blog-cacm/249538- by-nc-sa/3.0/
konrad-zuses-guestbook-a-treasure-trove/ [21] Deutsches Museum, “Patentstreit um Patent-
fulltext Anmeldung: Z 391 IXb /42m (Patent application dispute
[8] H. Bruderer, Milestones in Analog and Digital Z 391 IXb / 42 m),” parts 1-4, Archive IDs: NL 207/0980,
Computing, vol. 2, 3rd ed., Berlin, Germany: Springer NL 207/0981, NL 207/0982 and NL 207/0986, (browse
2020. [Online]. Available: https://www.springer.com/de/ using the arrows beneath the images): Part1: https://
book/9783030409739 digital.deutsches-museum.de/de/digital-catalogue/
[9] H. Bruderer, “Preparation for building the first Swiss archive-item/NL%2520207%252F0980/ Part2: https://
digital computer (Ermeth),” BLOG@CACM, Commun. digital.deutsches-museum.de/de/digital-catalogue/
ACM, Aug. 2020. [Online]. Available: https://cacm. archive-item/NL%2520207%252F0981/ Part3: https://
acm.org/blogs/blog-cacm/247088-preparation-for- digital.deutsches-museum.de/de/digital-catalogue/
building-the-first-swiss-digital-computer-ermeth/ archive-item/NL%2520207%252F0982/ Part4: https://
fulltext digital.deutsches-museum.de/de/digital-catalogue/
[10] H. Bruderer, “Why is there no well-known archive-item/NL%2520207%252F0986/
Swiss IT industry?,” BLOG@CACM, Commun. [22] I. Dey and D. Buck, The Cryotron Files. New York, NY,
ACM, Jan. 2020. [Online]. Available: https:// USA: The Overlook Press, 2018.
cacm.acm.org/blogs/blog-cacm/242462-why- [23] M. d’Ocagne, Le Calcul Simplifié; par les Procédés
is-there-no-well-known-swiss-it-industry/ Mécaniques et Graphiques, 3rd ed. Paris, France:
fulltext Gauthier-Villars, 1928.

July-September 2024 IEEE Annals of the History of Computing 33

ARTICLE

[24] J. P. Eckert and J. W. Mauchly, “Electronic numerical [35] P. E. Ludgate, “On a proposed analytical machine,” Sci.
integrator and computer (ENIAC),” U.S. Patent Proc. Roy. Dublin Soc., vol. 12, no. 9, pp. 77–91, Apr. 1909,
No.US3120606, filed Jun. 26, 1947, granted Feb. 4, 1964. see [55].
[25] K. Ganzhorn, “Historische Entwicklung der [36] E. Martin, Die Rechenmaschine und ihre
Informationsverarbeitung (Historic development of Entwicklungsgeschichte. Pappenheim, Germany:
data processing),” IBM Nachrichten, vol. 14, no. 164, pp. Johannes Meyer, 1925 (“Rechenmaschinen” for later
2152–2156, Feb. 1964. editions).
[26] W. Hoffmann, Ed. Digitale Informationswandler. Berlin, [37] U. C. Merzbach, “Zuse Interview (pp. 1–68 German,
Germany: Springer, 1962. English pp. 69–125, German pp. 126–154),” Deutsches
[27] IBM (New York) patent application, Museum, 1968. [Online]. Available: https://digital.
Datenverarbeitende Maschine mit sich zeitlich deutsches-museum.de/item/NL-207-0680/
überlappender Datenverarbeitung (Data processing [38] R. Messenger, “Triumph portable typewriters.” 2024.
machine with temporally overlapping data processing), [Online]. Available: https://oztypewriter.blogspot.
I17382 IXc/42m, filed Dec. 16, 1959, published Oct. 8, com/2012/05/triumph-portable-typewriters.html
1964, patent agent G.B. Hagen. [39] F. Nebeker, “Oral-history: Konrad Zuse,” interview
[28] IBM (New York) patent application, Maschine zur of Konrad Zuse, Interview #224, Engineering and
serienmäßigen Verarbeitung von Daten in binärer Technology History Wiki, IEEE Center for the History
Zeichenverschlüsselung (Machine for serial of Electrical Engineering, Aug. 28, 1994. [Online].
processing of data in binary character encryption), Available: https://ethw.org/Oral-History:Konrad_Zuse
J18712 IXc/42m, filed Sep. 14, 1960, published Jun. 4, [40] New York Times, “A basic computer patent at Sperry
1964, patent agent G.B.Hagen. Rand annulled,” Apr. 13, 1973. [Online]. Available:
[29] IEE Science Abstracts: Section A.—Physics, Institution https://www.nytimes.com/1973/04/13/archives/a-
of Electrical Engineers, in association with The basic-computer-patent-at-sperry-rand-annulledd-
Physical Society of London, The American Physical reason-for-ruling.html
Society, The American Institute of Electrical Engineers, [41] H. Petzold, “Die Mühlen des Patentamts,” in Die
and the Associazione Elettrotecnica Italiana, London, Rechenmaschinen von Konrad Zuse, R. Rojas, Ed.
Vol. XII, E. and F. N. Spon, London, Jun. 1909. Berlin, Germany: Springer, 1998, pp. 63–108.
[30] Konrad Zuse Internet Archive Home Page, 2024. [42] B. Randell, “Ludgate’s analytical machine of 1909,”
[Online]. Available: http://zuse.zib.de Comput. J., vol. 14, no. 3, pp. 317–326, 1971.
[31] Konrad Zuse Internet Archive, “Offizielle Patentakte [43] B. Randell, Ed. The Origins of Digital Computers:
von Telefunken, Patentabteilung. Gesamter Selected Papers, 1st ed. Berlin, Germany: Springer,
Schriftwechsel der Anmeldung Z391 Ixc/42m (Official 1973. Also 3rd ed., Heidelberg, Germany: Springer, 1982.
patent files of Telefunken, Patent Department. Entire [44] Rochester Democrat and Chronicle, “Who invented the
correspondence of application Z391 Ixc/42m),” computer,” p. 5C, Jan. 20, 1974.
filename zuse_archive-0533scan.pdf, archive ID [45] R. Rojas, Ed., Die Rechenmaschinen von Konrad Zuse.
0533, GMD No.006/001, 1966. [Online]. Available: Berlin, Germany: Springer, 1998.
http://zuse.zib.de/collection/BewakKVf30YcY53B/ [46] R. Rojas, Konrad Zuse’s Early Computers: The Quest
item/6CUfMX2RuO0KpmUo for the Computer in Germany (History of Computing).
[32] Lemelson Centre, “Honeywell vs. Sperry Rand records, Berlin, Germany: Springer, Oct. 2023.
1935-1973,” Smithsonian Institution, based on records [47] R. Rojas, “Konrad Zuse’s legacy: The architecture of the
held in the Hagley Museum & Library, Wilmington. Z1 and Z3,” IEEE Ann. Hist. Comput., vol. 19, no. 2, pp.
2024. [Online]. Available: https://invention.si.edu/ 5–16, Apr.–Jun. 1997.
honeywell-vs-sperry-rand-records-1935-1973 [48] R. Rojas and U. Hashagen, “The architecture of
[33] A. Lovelace, “Sketch of the analytical engine invented by Konrad Zuse’s early computing machines,” in The First
Charles Babbage, esq. By L.F. Menabrea, of Turin, officer Computers—History and Architectures, R. Rojas and H.
of the military engineers, with notes by the translator,” Hashagen, Eds., Cambridge, MA, USA: MIT Press, 2000,
in Scientific Memoirs, vol. 3. R. and J. E. Taylor, Eds, pp. 236–261.
London, U.K.: Taylor and Francis, 1843, pp. 666–731. [49] C. E. Shannon, “A symbolic analysis of relay and
[34] P. E. Ludgate, “Automatic calculating machines,” in switching circuits,” Trans. Amer. Inst. Elect. Eng., vol. 57,
Napier Tercentenary Celebration: Handbook of the pp. 471–495, 1938.
Exhibition (Also Published as Modern Instruments and [50] J. Smiley, The Man who Invented the Computer: The
Methods of Calculation), E. M. Horsburgh, Ed., London, Biography of John Atanasoff. New York, NY, USA:
U.K.: G. Bell and Sons Ltd., 1914, pp. 124–127, see [55]. Doubleday, 2011.

34 IEEE Annals of the History of Computing July-September 2024

 ARTICLE

[51] A. P. Speiser, The Early Years of the Institute: [59] G. P. Zachary, Endless Frontier: Vannevar Bush,
Acquisition and Operation of the Z4, Planning of the Engineer of the American Century. New York, NY, USA:
ERMETH, Dept. Comput. Sci., ETH Zürich, Zurich, Free Press, 1997.
Switzerland, 1998. [60] K. Zuse, Rechenvorrichtung (Calculating device),
[52] The Charles Babbage Institute, “ENIAC-the first Patent Application Z26476 (Z391) IXb/42m, German
computer?,” University of Minnesota, Minneapolis, Patent Office, filed Jun. 16, 1941, published Dec. 4, 1952.
MN, USA, 2024. [Online]. Available: https://gallery.lib. [61] K. Zuse, The Computer—My Life. Berlin, Germany:
umn.edu/exhibits/show/digital-state/eniac Springer, 1993.
[53] I. Travis, “Bibliography of literature on calculating
machines,” Moore School Electrical Engineering, Univ.
BRIAN COGHLAN is a retired Senior Lecturer with the School
Pennsylvania, Philadelphia, PA, USA, 1938.
[54] I. Travis, “Statement in evidence to the ENIAC trial,” of Computer Science and Statistics, Trinity College Dublin, The
IRVEN TRAVIS 6542, p. 14, Trial Transcript III 33, University of Dublin, D02 X9W9, Dublin, Ireland. He is curator of
testimony given on 11th August 1971. The John Gabriel Byrne Computer Science Collection. Contact
[55] Trinity College Dublin, “Investigating Percy Ludgate him at [email protected].
and his Analytical Machine (SCSS Ludgate webpage),”
The John Gabriel Byrne Computer Science Collection,
BRIAN RANDELL is an Emeritus Professor of Computer Science
2024. [Online]. Available: https://www.scss.tcd.ie/
SCSSTreasuresCatalog/ludgate/ with the School of Computing, Newcastle University, NE4 5TG,
[56] Trinity College Dublin, “Index of SCSS Ludgate-Zuse Newcastle upon Tyne, U.K. His book The Origins of Digital
Files,” The John Gabriel Byrne Computer Science Computers: Selected Papers was first published in 1973, by
Collection, 2024. [Online]. Available: https://www. Springer. Contact him at [email protected].
scss.tcd.ie/SCSSTreasuresCatalog/miscellany/
TCD-SCSS-X.20121208.002/ZuseFiles/
RALF BUELOW is a former museum curator and a free-lance
[57] Wikipedia, “Honeywell, Inc. v. Sperry Rand Corp,” 2024.
[Online]. Available: https://en.wikipedia.org/wiki/ historian of technology and science journalist. He is based
Honeywell,_Inc._v._Sperry_Rand_Corp in Berlin and works for the Heinz Nixdorf MuseumsForum,
[58] M. V. Wilkes, Automatic Digital Computers. London, 33102, Paderborn, Germany. Contact him at ralf.buelow@
U.K.: Methuen, 1956. t-online.de.

IEEE Computer Society

Has You Covered!
WORLD-CLASS CONFERENCES —
Over 195 globally recognized conferences.
DIGITAL LIBRARY — Over 900k articles covering
world-class peer-reviewed content.
CALLS FOR PAPERS — Write and present your
ground-breaking accomplishments.
EDUCATION — Strengthen your resume with the
IEEE Computer Society Course Catalog.
ADVANCE YOUR CAREER — Search new positions
in the IEEE Computer Society Jobs Board.
NETWORK — Make connections in local Region,
Section, and Chapter activities.

Explore all member benefits

www.computer.org today!