Appendix

World-wide Developments

We have in the body of the text discussed the world-wide developments that directly affected progress in the computer field. There were however many others that had extremely important local effects, and we discuss or at least sketch them here.

GREAT BRITAIN. As we saw, the British were very active from the beginning and produced a variety of excellent machines (above, pp. 217–219, 246–249). We need say no more about their activities, except to say a few words more about Andrew D. Booth. He was in those days—1947—at Birkbeck College, London University, working on relay calculators. His ARC, Automatic Relay Computer, was his first machine and was completed in cooperation with the present Mrs, Booth, Kathleen H. V. Britten. This was followed by a very small machine SEC, Simple Electronic Computer, and then by a series of APEC’s, All Purpose Electronic Computers, built for Birkbeck for x-ray crystallography and for a number of other users. These were built in the 1949–1951 era and were at least partly the results of Booth’s and Miss Britten’s year at the Institute for Advanced Study in 1946–47, a visit made possible by a grant from the Rockefeller Foundation. The Booths were clearly somewhat ahead of their time in some respects; as mentioned earlier, he conceived of the magnetic core and lectured on it at about the same time as did Forrester.

SWEDEN. The Swedish Government, as we mentioned (above, pp. 249ff), sent Stig Ekelof to the United States in mid-1946 to look into the computing field. In November or December 1948 Sweden had appointed a five-man board headed for a short while by Rear-Admiral Stig Hanson-Ericson and then in April 1950 by the Under-Secretary of Defense, Gustav Adolf Widell, as chairman, with Gunnar Berggren, Ekelöf, Carl-Erik Fröberg, and Commodore Sigurd Lagerman as advisors. In December 1948 the Board authorized a highly talented man, Conny Palm, to head “a Working Group” to do research and development work on machines. The first accomplishment of the group was a relay machine.1 Palm and Gösta Neovius built the BARK, Binar Automatic Rela-Kalkylator, at the Royal Institute of Technology in Stockholm with funding by the Swedish Telegraph Administration. They were supported by Fröberg, G. Kjellberg, and a number of other talented people. The design of the machine was due to Neovius and Harry Freese. It was completed in February 1950, ²

Fröberg, Kjellberg, and Neovius were members of the group who spent the academic year 1947–48 in the United States. The group consisted of four first-rate men, both physicists and engineers: Carl-Erik Fröberg, now Professor at Lund University and head of its Institute of Computer Sciences, spent his time at the Institute for Advanced Study, where we became good friends. He is both a distinguished and delightful person, with a fine sense of humor. Goran Kjellberg spent the year with Howard Aiken’s people at Harvard; Gösta Neovius was at the Massachusetts Institute of Technology; and Erik Stemme, now Professor at the Chalmers Institute of Technology in Gothenberg, visited at RCA, Princeton and then with us at the Institute for Advanced Study.

Stemme became chief engineer of a project to build an electronic computer. The Working Group under Stemme then built the BESK, Binär Electronisk Sekvens Kalkylator, at the Royal Institute along the lines of the Institute for Advanced Study machine. It was completed in November 1953, and then in September 1956 its Williams tube memory was replaced by a 1, 024-word magnetic core memory.

Fröberg built his SMIL, Siffermaskinen I Lund, as a magnetic drum machine in organization like BESK, and hence like the Institute for Advanced Study machine. It was put into operation in June 1956 and decommissioned on February 3, 1970.³ This machine, as well as BESK, played an important role in training of Swedish students in computer engineering and science.

Since those times the computing people in Sweden have turned their attention with great success to numerical analysis and the computer sciences away from machine construction.⁴

DENMARK. In Denmark there was also considerable interest in computing due largely to one man, Professor Richard Petersen of the Technical University in Copenhagen. He was a student, close personal friend, and mathematical collaborator of Harald Bohr from 1933 until Bohr's untimely death. Petersen in addition to being a fine mathematician was a truly fine person of tremendous integrity, goodwill and sincerity.5

As early as 1946 the Danish Academy of Technical Sciences formed a committee chaired by Petersen and containing a number of leading figures in Denmark with an interest in computation. Among others, the committee contained Niels E. Nørlund and Bengt G. D. Strömgren. The former was Niels Bohr’s brother-in-law and is the author of many fundamental papers on finite difference methods and on geodesy.⁶ I first met him at Petersen’s home about fifteen years ago and had the pleasure of seeing the beautiful maps he had made in his laboratory, the Danish Geodetic Institute.⁷ Bengt Strömgren is now professor of astrophysics at the University of Copenhagen, president of the Royal Danish Academy of Sciences and Letters and of the International Astronomical Union. He has had a most distinguished career both in Europe and in the United States, where he was a distinguished service professor at the University of Chicago and later a professor at the Institute for Advanced Study.

In any case the Danish committee started construction of both a differential analyzer and an electronic equation solver. Then came the UNESCO proposal, and Richard Petersen had occasion to meet and talk with American and European colleagues in Paris. These talks helped convince him of the desirability of electronic digital machines. He thereupon applied to the Carlsberg Foundation for financial support. Fortunately the Board of Directors of this very enlightened foundation, which included Prof. Børge Jessen, the distinguished mathematician, was very helpful, and Petersen set up the Regnecentralen—Danish Institute of Computing Machinery—as a division of the Danish Academy of Technical Sciences. He also secured the blessing of the Swedish Board to make a modified copy of the BESK. In particular, the intention was to use a core memory. This machine was dedicated in the summer of 1957, Fortuitously, my family and I arrived in Copenhagen in time to take some small part in the dedication festivities of the DASK, the first machine produced by the Regnecentralen.8

NORWAY. The Norwegian government was much less active than were the Swedish or Danish ones. But it did send Dr. Ernest S. Selmer of the University of Oslo to the Institute for Advanced Study for the second terms of 1950–51 and 1951–52. His university then procured APE(x)c from Booth. It also built a small machine called NUSSE that was described by the Nobel Laureate in economics, Ragnar Frisch.⁹

THE NETHERLANDS. The computing activity in the Netherlands started on 11 February 1946 with the establishment of the Mathe-matisch Centrum. When started it consisted of four departments headed by distinguished scientists: pure mathematics under J. G. van der Corput and F. J. Koksma; mathematical statistics under D. van Dantzig; applied mathematics under B. L. van der Waerden; and computation under A. van Wijngaarden. The latter is one of its most vigorous members and has become a leader in the ALGOL community. He first built a relay machine in 1948–1951 called ARRA. Then a vacuum tube version with a magnetic drum was built and finished in 1954 by van Wijngaarden. This was in turn followed by ARM AC, Automatische Rekenmachine Mathematisch Centrum, in June 1956.¹⁰

It would be very wrong not to mention van Wijngaarden’s colleagues: G. Blaauw, B. J. Loopstra and C. S. Scholten. Another Hollander of this period is W. L. van der Poel who independently built a small machine at the Postal Telephone and Telegraph Laboratory in the Hague called PTERA; he has long played a significant role in the programming field.11

FRANCE. We have already mentioned Louis Couffignal’s activities at the Institute Blaise Pascal. In addition the Compagnie des Machines BULL brought out in 1956 its GAMMA 3 ET, a small electronic machine.¹² Quite soon after the war an organization known as SEA, Société d’Electronique et d’Automatisme, was formed by F. H. Raymond, an exceedingly good engineer. This organization built a series of electronic computers under the generic name of CAB, Calculatrice Arithmetique Binaire.

SWITZERLAND. We have already mentioned Konrad Zuse and his key position in the German computing field. His Z4 was the starting point for the Swiss at the ETH in the year 1950. There Edward L. Stiefel founded an Applied Mathematics Institute and gave life to a whole school of excellent numerical analysts who have made Zurich prominent in the field.¹³

Another man from this Institute is a first-class engineer, Ambros Speiser, who received his engineering degrees from the ETH in 1948 and 1950. He spent the year 1948–49 working in Aiken’s laboratory at Harvard preparing himself to lead the engineering staff at the ETH in building the ERMETH, Electronische Rechenmaschine der Eidgenossischen Technischen Hochschule. The machine was completed in 1955 and showed the influences both of Zuse’s Z4 and Aiken’s Mark IV.14 Speiser went from there to distinguish himself further by becoming the first head of IBM’s Research Laboratory in Zurich and then Director of Research for the well-known Swiss engineering firm of Brown, Boveri, where he is now.

From our point of view one of the most important figures in Zurich was Heinz Rutishauser, who played a key role in the development of automatic programming from its earliest days on through ALGOL 60 (above, p. 337).

GERMANY. In West Germany concern for electronic computers developed quite rapidly and in several places. Just as in the United States, World War II triggered off considerable interest in Germany in computing in general and in electronic computers in particular. There are two engineers who filed patent applications which show they had at a very early date the concept of using vacuum tubes to expedite computation. Walter Hiindorf of Munich conceived of his Elektrische Rechenzelle on 6 April 1959. Helmut Schreyer, a colleague of Konrad Zuse (above, p. 250), conceived of his device on 11 June 1943, although he apparently started developing his ideas as early as 1937. Hiindorf’s ideas seem to have centered around special computing vacuum tubes and appear to me to be similar to comparable inventions of J. A. Rajchman, Richard Snyder, and others at RCA made during the wartime era. All these special tubes, in the event, disappeared in favor of simple, general-purpose tubes linked together by external circuitry. The special-purpose tubes were extremely difficult to make with the glass seal technologies of that time, and because of their specialized nature—hence low production volume—they were very expensive.

Schreyer’s work was done at the Technische Hochschule in Berlin as his thesis topic. A small piece was constructed, but the machine itself was shelved by the Nazis as being “völlig irreal und unwichtig” ¹⁵ I suspect—although I do not know—that de Beau-clair’s claims for Schreyer’s machine are overenthusiastic.

Perhaps the first significant developments in Germany arose at the Max Planck Institut fü Physik in Göttingen. This work was in large measure due to Heinz Billing in collaboration with Ludwig F. B. Biermann, the astrophysicist. Their first machine, Gl, was started in 1950, finished during 1951, and put in regular operation in 1952.16 This machine was more or less a test model for G2 (there was also a Gla)—which was in normal operation by December 1954 and not 1959 as stated by de Beauclair. Both Gl and G2 were serial machines. They were followed by the G3, a parallel machine.¹⁷ (We at the Institute for Advanced Study were fortunate in having Dr. Elenore Trefftz of Billing’s staff with us during the academic year 1952–53 and Billing in about 1956.) Later, in 1958, the Max Planck Institut was moved by its Director Werner Heisenberg to Munich, and it expanded first into the Max Planck Institut für Physik und Astrophysik and eventually into two separate institutes, one for physics under Heisenberg and one for astrophysics under Biermann. During this era a number of extremely able young astrophysicists from there such as R. H. F. Lust and A. Schliiter came to Princeton University to study with Martin Schwarzschild. While in Princeton they did a number of interesting and important calculations on the computer at the Institute for Advanced Study.

The development of computers in Germany was not confined to the Institut für Astrophysik. Already in 1952 Professor Hans Piloty of the Technische Hochschule of Munich, together with his son Robert, now Professor at the Technische Hochschule of Darmstadt and then a privatdozent who had recently received his doctorate at MIT, were very busy developing their machine. Others in this project should also be mentioned since they also have become important figures in the computer field. They are Walter Proebster and Hans-Otto Leilich of Piloty’s Institut fü Elektrische Nach-richtentechnik und Massteclmik as well as Friedrich L. Bauer and Klaus Samelson of the Matheinatische Institut.¹⁸

This machine was known as PERM, Programmgesteurte Elek-tronische Rechenanlage Munehen, and was very fast; it had both a core memory of 2, 048 words and a magnetic drum of 8, 192 words. Piloty very charmingly—from our point of view—described his machine in these words: “We took the famous work of Professor v. Neumann and Professor Goldstine as model and guide for our first considerations. Since Dr. Goldstine is present here on our symposium, I am very glad of being able to express my deep gratitude to our teachers before him personally.”19

Another early development in West Germany took place in Darmstadt where Professor Alwin Walther (1898–1967) headed the Institut fü Praktische Mathematik at the Technische Hochschule. He was engaged in ballistic work for the military in Germany, and in 1943 or 1944 already began to conceive of a digital computer to aid his work. Then in 1951, with the help of his colleagues Hermann Bottenbush, Hans-Joachim Dreyer, Walter Hoffmann, Walter Schiitte, Heinz Unger, and others, Walther began to build his DERA, Darmstadter elektronische Rechenautomat. It was not completed however until 1959.²⁰ The machine designer was much influenced by the ideas of Aiken, whom Walther admired greatly. In speed the machine was rather slow, a multiplication taking about 12 milliseconds and an addition about 0.8 milliseconds in contrast to the PERM which did an addition in about 9 microseconds.

These developments in the West German universities plus those of Zuse in industry served to stimulate a considerable interest in the computer field. At the present, this is so great that the Federal Government is putting substantial amounts of money into the Lander operated universities in order to encourage the creation of computer science departments.

In East Germany there was considerably less progress in electronic machines except at Dresden, where N. Joachim Lehmann built the Dl (completed in 1956) and its extension, the D2, at the Institute of Technology.²¹

AUSTRIA. During the 1950s the Institut für Niederfrequenztechnik of the Institute of Technology in Vienna was as active as its finances permitted. Under the direction of Heinz Zemanek, a distinguished leader of computer science in Europe, this Institute used an IBM 604 to solve problems; it built a very small—about 700 relays—relay computer called URR-l, Universalrechenma-schine 1; a small relay machine to solve logistical problems LRR-1; and in the period 1955–1958 a transistorized machine with a magnetic drum.22 Its multiplication time was around 0.4 milliseconds, and it was known as a little May breeze, Mailüfterl. In an address on the machine, Zemanek with his usual Viennese charm compared the Austrian efforts with those of the Americans and said: “… wenn die Amerikaner einen ‘Whirlwind’ besitzen dann miisste man dieser Maschine auf gut wienerisch den Namen ‘Mailüfterl’ geben.”

ITALY. The Italians did not undertake the development of an electronic computer with the speed of some other European countries, but they nonetheless set up a group at Pisa with the idea of learning the field. This resulted in the CEP, Calcolatrice Elettronica Pisana, an asychronous machine with cores and drums, produced after the time our history covers. The Institute at Pisa was known as Centro Studi Calcolatriei Elettroniche. Its machine was in some ways similar to the Institute for Advanced Study one in logical design. The International Center used commercially available equipment as did Professor Mauro Picone’s famous Istituto Nazion-ale per le Applicazioni de Calcolo. This institute was then (1957) about 25 years old and had trained an extremely large group of applied mathematicians.

BELGIUM. Professor C. Manneback of Louvain University in Belgium was one of several scientists who was, from an early time, interested in the development of computers. In part, his interest caused the Institut pour TEncouragement de la Recherche Scien-tifique dans lTndustrie et P Agriculture and the Fonds National de la Recherche Scientifique to commission the Bell Telephone Manufacturing Co. of Antwerp to build a machine for Belgium. The construction started in 1951 and was finished in 1954. It was a magnetic drum machine with a 16-mi Hi second multiplication time and was used at the Centre d’Étude et d’Exploitation des Calculatrices Electronique in Brussels.23 The design reflected the profound influence Aiken had on European computer design.²⁴

RUSSIA. The Russians have been very concerned about electronic computers from the late 1940s on. I had on a number of occasions received requests from a Russian trading company for the reports by von Neumann and myself on electronic computing instruments. By 1953 the BESM, Bystrodeistwujuschtschaja Elektronnajastschet-naja Machina, one of the first Russian machines, was completed. In 1955 it had a Williams tube memory of 1, 024 words and a magnetic drum of 5, 120 words arranged in five groups of 1, 024 each. It also had a small—376 words—germanium diode memory. The operating times were quite good—between 77 and 182 microseconds for addition and 270 for multiplication. Later the Williams tubes were replaced by magnetic cores.

This machine was designed under the direction of Academician Sergei A. Lebedev of the Academy of Sciences of the USSR, Moscow. It was announced to the western world at the Darmstadt conference in 1955.²⁵ This Russian computer has a 39-binary-digit word together with a three-address code and uses a floating point. It was developed at the Institute of Precise Machines and Computing Techniques.

At the same conference I.I. Basilewski announced and discussed the URAL, a magnetic drum computer completed in 1955, It was built at the Scientific Research Institute of the Ministry of Machine and Instrument Construction under the direction of B. I. Rameev. It has 36 bit words, a 1, 024-word drum, and a multiplying speed of about 10 milliseconds. This machine is said to be the prototype for a series of over 300 copies.²⁶

The STRELA was another large Williams tube machine with a capacity of 1, 023 words each of 43 binary digits. It, like BESM, was a floating point machine and like it has magnetic tapes, which can hold about 200, 000 words. This machine was constructed in 1953 under Basilewski’s direction.27 Later the Williams tubes were replaced by cores and about 15 copies were made by 1960. Its multiplication time was about 500 microseconds.

In addition to these three a number of other computers were built in the USSR during the same period: the PAGODA, the Ml and M2, MESM, KRISTALL, N.12. Moreover, since 1957, our cutoff date, a number of others have been started. The interested reader may wish to consult a paper by Carr, Perlis, Robertson, and Scott describing their fortnight in the USSR in the late summer of 1958.²⁸ They stated in their conclusions the following germane comparisons: “As to the equipment seen, the British EDSAC II (Cambridge) and Danish DASK were of later design than the BESM I. All three are of comparable speed.”

CZECHOSLOVAKIA. In this country Antonin Svoboda was a pioneer. As early as 1952 he had been involved in building various kinds of calculating devices, and at the Darmstadt conference of 1955 he described his ARITMA as having “interesting similarities with the DERA computer.”²⁹ (ARITMA is apparently a nationalized company making punch card equipment.) In addition, Svoboda and his colleagues at the Institute of Mathematical Machines of the Academy of Science built several small machines and a larger relay machine SAPO, Samoinný Poíta, with a magnetic drum.³⁰

Svoboda is quite well known for his work on so-called residual classes—a very interesting and novel way to do arithmetic. This work was pioneered by M. Valach, who discovered the procedure, and Svoboda.

JUGOSLAVIA. The first machine in Jugoslavia was to be a magnetic core machine and was not due to be completed during the span of this account.

POLAND. In Poland three machines were made, two of them at the Instytut Maszyn Mathematysznych in Warsaw and one at the Research Institute for Electronic Computers of the Polish Academy of Sciences.³¹

JAPAN. The Japanese were deeply concerned about computers and very early recognized their importance to their country. About 1943 an analogue computer was built to solve systems of linear equations. This work was done in the Electrotechnical Laboratory of the Ministry of Communications. In the same period a Bush-type differential analyser was installed in the Aeronautical Laboratory of Tokyo University. Then in 1944 the Japanese Scientific Research Council set up a research committee on electrical computing machines with Hideo Yamashita as chairman. This committee seems to have been active in encouraging the use of differential analysers, since six were installed in the period 1944–1952, one of which used electronic techniques.32

In 1951 the Japanese Permanent Delegate to UNESCO, Mr. Hagiwara, was already quite anxious to have a sub-center of the proposed International Computational Centre established in Japan. He and Yamashita met for a number of discussions with me in Paris in November 1951, and they followed this with a visit of some of their colleagues a few years later.

In one of the articles mentioned in note 32 there is a reference to the TAC, Tokyo Automatic Computer. It says that the device was to be completed in 1952—then, in handwriting, “(if finances be obtained).” Apparently they were eventually, and the machine was completed in 1956 for the Research Committee for Electric Computers of the National Research Council of Japan in Tokyo. It used an electrostatic storage tube memory of 512 words and a magnetic drum of 1, 536 words. It used binary numbers and a single-address code. It did a multiplication in about 8 milliseconds.

This was not the first machine the Japanese built. That was the so-called E.T.L. Mark I completed in 1952, which was followed by the E.T.L. Mark II in November 1955.³³ They were built under the direction of Motinori Goto with Yasuo Komamiya as chief designer.³⁴ These are relay machines. (Later, in 1957, Goto was to go to Harvard to study for a year in Aiken’s laboratory under a United Nations Scholarship.) Then came E.T.L. Mark III and IV, both of which were fully transistorized computers.35 It is really remarkable that the Japanese could have finished a relay computer as late as November 1955 and a transistorized one in November 1957.

In 1954 Eiichi Goto invented an ingenious device called a parametron. At the same time, von Neumann independently conceived of the same device. “However von Neumann’s proposal has not been followed up by the computer designers nor brought into practice until the successful Japanese developments in this field became known.”³⁶

ISRAEL. The Israelis were also active in our field from an early stage. Chaim L. Pekeris, who has headed the applied mathematical work at the Weizmann Institute of Science since its inception, was at the Institute for Advanced Study on a number of occasions; his first visit was from 1946 through 1948 when he became very interested in having an electronic computer at his institute. He sent E. Frei there and later J. Gillis and also persuaded Gerald Estrin to take a leave of absence from the Institute for Advanced Study to build the WEIZAC.³⁷ It was started in June 1954 and tests were to begin in March 1955. The machine used a drum memory. It was a member of the Institute for Advanced Study family.

RUMANIA. The Institute of Physics of the Rumanian Academy of Sciences built several machines for itself and the University of Bucharest known as CIFA-1, 2, 3.³⁸

INDIA. Finally, the well-known Tata Institute of Fundamental Research in Bombay, India, started early in 1955 the development of a pilot of an electronic machine of the Institute for Advanced Study type. It was to have a core memory of 256 words and was in November 1956 doing a number of test calculations.39

¹ Much of the material here is contained in a few mimeographed bulletins issued by the Board in its early days.

² G. Kjellberg and G. Neovius, “The BARK, A Swedish General Purpose Relay Computer, ” MTAC, vol. 5 (1951), pp. 29–34.

³SMIL means “smile” and BESK is the slang word for “beer” and means “bitter.”

⁴ Among the distinguished numerical analysts who were involved from the earliest days with the Working Group, under the direction first of Palm (who died prematurely in December 1951), Neovius, and Stig Comét, and then Gunnar Havermark, are Carl-Erik Fröberg, O. Karlqvist, and Germund Dahlquist, now Professor at the Royal Institute of Technology in Stockholm, who was head of the mathematical staff in the Working Group from 1 May 1956. He today is a very important numerical analyst with a worldwide reputation. It is indeed interesting how all these people in Sweden are carrying on a century-old tradition dating from Seheutz’s time.

⁵ In a touching tribute to him his biographer has truly written: “His friends, his colleagues, and the thousands of students whom he had the opportunity of teaching, retain innumerable valuable and merry rocollections of a man in perfcct mental equilibrium, always on the move towards new goals, always ready to help, and thoroughly enjoying every task which he took upon himself.” Erik Flansen, Obituary of Professor Dr. Richard Petcarsen, tr;uislated by James Steffensen and privately distributed by Mrs. Kale11 Richard Petersen.

⁶ K.g., N. E. Nørll.rntl, Vorlesungen Über Dz:ferenzenwchung (Berlin, 1924).

⁷ During World War II the Nazis directed this institute to work for them. Nørlund managed to persuade thrwi that he bc allowed to prepare maps of medieval Danish ports and harbours! Thesc maps were of course no nse whatsoever to the Nazis but are of great impnrtancbe to scholars.

⁸ This organization subsequently became a commercial venture. Petersen’s biographer has noted that: “He doubtless considered his work in connection with the institute to be an extension of his work at the Technical University, and he did not conceal his regret that the hopes of a closer connection between these two institutions were not fulfilled.”

⁹ R. Frisch, Notes on the Main Organs and Operation Technique of the Oslo Electronic Computer NUSSE, Memorandum of the Socio-Economic Institute of the University of Oslo, October 1956.

¹⁰ A brief résumé of the machine’s characteristics may be found in Journal of the ACM, vol. 4 (1957), pp. 106–108. This is a portion of an Office of Naval Research Digital Computer Newsletter which was reprinted by ACM. See also, A. van Wijngaarden, “Moderne Rechenautomaten in den Niederlanden, ” Nachrichten-technische Fachberichte (Braunschweig), vol. 4 (1956), pp. 60–61. Fortunately a big and important symposium on electronic computers was held in Darmstadt on 25–27 October 1955. At this meeting virtually all European machine developers presented papers. The proceedings appeared in the issue of Nachrichtentechnische Fachberichte just cited; hereafter referred to as N.F.

¹¹ W. Id. van der Poel, “The Essential Types of Operations in an Automatic Computer, ” N.F., pp. 144–145; B. J. Loopstra, “Processing of Fortnulas by Machines, ” N.F., pp. 146–147; C. S. Scholten, “Transfer Facilities hrtween Momories of Different Types, ” N.F., pp. 118–1 19.

¹² See W. de Beauclair, Recltnen mit Maschinen, pp. 184–185, for a discussion of other French machines in thc post-1956 era.

¹³ The list includes, among others, Pttter Henrici, Urs Hochstrasser, Werner Leiltett, Werner Liniger, Hans Maehly, Heirlz Rutishauser, and Andreas Schopf.

¹⁴ A, P. Speiser, “Eingangs-und Ausgangsorgane sowie Schaltpulkte der ERMETH, ” N.F., pp. 87–89.

¹⁵ de Beauelair, Rechnen mit Maschinen, p. 206.

¹⁶ L. Biermann, “Überblick üiber die Göttinger Entwieklungen, insbesondere die Anwendung der Maschinen Gl und G2, ” IV.F., pp. 36–39; II, Öhlmann, “Bericht uber die Fertigstellung der G2, ” N.F., pp. 97–98; and K. Pisula, “Die Weiterent-wicklung des Befehlscodes der G2, ” N.F., pp. 165–167.

¹⁷ A. Schliiter, “Das Göttinger Projekt einer Schnelten Elektronischen Reehen-maschine (G3) “ N.F., pp. 99–101.

¹⁸ H. Piloty, “Die Entwicklung der Perm, ” N.F., pp. 40–45; W. E. Proebster, “Dezimal-Binäar-Konvertierung mit Gleitendem Komma, ” N.F., pp. 120–122; F. L. Bauer, “Interationsverfahren der Iinearen Algebra von Bernoullischen und Graef-feschen Korvergenztyp, ” N.F., pp. 171–176; and R. Piloty, “Betrachtungen iiber das Problem der Datenverarbeitung, ” N.F., pp. 5–8.

¹⁹ H. Piloty, op. cit.

²⁰ H. J. Dreyer, “Der Darmstädter elektronische Rechenautomat, ” N.F., pp. 51–55; W. Schütte, “Einige technische Besonderheiten von DERA, N.F., pp. 126–128; H. Unger, “Arbeiten der Darmstädter mathematischen Rechenautomat, ” N.F., pp. 157160; H. Bottenbusch, “Unterprogramme für DERA, ” JV.F., pp. 165–167.

²¹ N. J. Lehmann, “Stand und Ziel des Dresdender Rechengeräte-Entwicklung, ” N.F., pp. 46–50; Lehmann, “Bericht über den Entwurf eines kleinen Rechenautomaten an der Technischen Hochschule Dresden, ” Ber. Math. Tagung, Humboldt-Universität Berlin (Berlin, 1953), pp. 262–270; also Lehmann, “Bermerkungen zur Automatisierung der Programmfertigung für Rechenautomaten (Zusammefassung), ” N.F., p. 154.

²² H. Zemanek, “Die Arbeiten an elektronischen Rechenmaschinen und Informa-tionsbearbeitungsmaschinen am Institut für Niederfrequenztechnik der Technischen Hochschule Wien, ” N.F., pp. 56–59. See also Zemanek, “‘Mailüfterl, ’ ein dezimaler Volltransistor-Rechenautomat, ” Elektrontechnik u. Maschinenhau, vol. 75 (1958), pp. 453–463.

²³ 1. L. F. de Kerf, “A Survey of European Digital Computers, ” Parts I, II, III, Computers and Automation, vol. 9 (1960).

²⁴ M. Linsman and W. Pouliart, “Principales Caractéristiques dans la Machine Mathématique IRSIA-FNRS, ” N.F., pp. 66–68, and V. Belevitch, “Le Trafic des Nom-bres et des Ordres dans la Machine IRSIA-FNRS, ” N.F., pp. 69–71.

²⁵ S. A. Lebedev, “BESM, eine schneliaufende elektronische Rechenmaschine der Akademie der Wissenschaften der USSR, ” N.F., pp. 76–79,

²⁶ I.I. Basilewski, ‘“Die universelle Elektronen-Rechenmaschine URAL für ingenieur-technische Untersuchungen, ” N.F., pp. 80–86.

²⁷ A. I. Kitov and N. A. Krinitskii, Electronic Computers, translated from the Russian by R. P. Froom (London, 1962).

²⁸ J. W. Carr III, A. J. Perlis, J. E. Robertson, and N. R. Scott, “A Visit to Computation Centers in the Soviet Union, ” Comm. of ACM, vol. 2, no. 6 (1959), pp. 8–20. This article also contains an excellent bibliography.

²⁹ A. Svoboda, “ARITMA Calculating Punch, ” N.F., p. 72.

³⁰ J.Oblonsky, “Some Features of the Czechoslovak Relay Computer SAPO, ” N.F., pp. 73–75.

³¹ de Beauclair, Rechnen mit Maschinen, p. 103.

³² Memorandum, The recent development of computing devices in Japan. This is in my files but is undated and unsigned but has attached to it reprints of two articles in Japanese describing an A. C. Network Analyzer and a Differential Analyzer. Both of these are dated October 1951. They were probably given me by Yamashita and his associates.

³³ M, Goto et al., “Theory and Structure of the Automatic Relay Computer, E.T.L. Mark II, ” Researches of the Electrotechnical Laboratory (Tokyo, 1956).

³⁴ W. Hoffmann, Digitale lnformationstoandler (Braunschweig, 1962). In particular, see M. Goto and Y. Komamiya, “The Relay Computer E.T.L. Mark II, ” pp. 580–594. This book is an excellent compendium of material on computers worldwide prior to 1962 and has truly remarkable bibliographies in it.

³⁵ Ibid., pp. 575–649, “Digital Computer Development in Japan.” This section of Hoffmann’s Digitate Informationswandler is a series of articles, with bibliography, under the editorship of H. Yamashita. The authors are M. Goto and Y. Komamiya, H. Takahasi and E. Goto, S. Takahashi and H. Nishino, T. Motooka, N. Kuroyanagi. The topics covered are the Mark U, the Parametron, Memory Systems for Parametron Computers, Mark IV, Magnetic Core Circuits, the Esaki Diode, and High-Speed Arithmetic Systems.

³⁶ Ibid., p. 595.

³⁷ Office of Naval Research, Digital Computer Newsletter, reprinted in Jour, of ACM, vol. 2 (1955), p. 135.

³⁸ V. Toma, “CIFA-1, the Electronic Computer of the Institute of Physics of the Academy of the Rumanian People’s Republic, ” Inter. Math-Koll., Dresden, 22–27 November 1957.

³⁹ Office of Naval Research, Digital Computer Newsletter, in Jour, of ACM, vol. 3 (1956), p. 110.