Konrad Zuse

Born June 22, 1910, Berlin-Wilmersdorf; German inventor of pre-war electromechanical binary computer designated Z1 which was destroyed without trace by wartime bombing; developed two more machines before the end of the war but was unable to convince the Nazi government to support his work; fled with the remains of Z3 to Zurich where he developed the Z4 which was successfully used at ETH. Developer of a basic programming system known as "Plankalkül" with which he designed a chess playing program. Education and Experience: By 1927 Konrad Zuse had enrolled at the Technical University in Berlin-Charlottenburg and began his working career as a design engineer (Statiker) in the aircraft industry (Henschel Flugzeugwerke) and by 1935 he had completed a degree in civil engineering. He remained in Berlin from the time he finished his degree until the end of the war in 1945, and it was during this time that he constructed his first digital computers. He later formed his own company for the construction and marketing of his designs.

During 1936 to 1938 Konrad Zuse developed and built the first binary digital computer in the world (Zl). A copy of this computer is on display in the Museum for Transport and Technology ("Museum fur Verkehr und Technik") (since 1989) in Berlin.

The first fully functional program-controlled electromechanical digital computer in the world (the Z3) was completed by Zuse in 1941, but was destroyed in 1944 during the war. Because of its historical importance, a copy was made in 1960 and put on display in the German Museum ("Deutsches Museum") in Munich.

Next came the more sophisticated Z4, which was the only Zuse Z-machine to survive the war. The Z4 was almost complete when, due to continued air raids, it was moved from Berlin to Gottingen where it was installed in the laboratory of the Aerodynamische Versuchanstalt (DVL/Experimental Aerodynamics Institute). It was only there for a few weeks before Gottingen was in danger of being captured and the machine was once again moved to a small village "Hinterstein" in the Allgau/Bavaria. Finally it was taken to Switzerland where it was installed in the ETH (Federal Polytechnical Institute/"Eidgenossisch Technische Hochschule") in Zurich in 1950. It was used in the Institute of Applied Mathematics at the ETH until 1955.

My first computer and first thoughts about data processing

I started in 1934, working independently and without knowledge of other developments going on around me. In fact, I hadn't even heard of Charles Babbage when I embarked on my work. At that time, the computing industry was limited to mechanical calculators using the decimal system. Punched card devices were slightly further developed and able to deal with relatively complex operations for statistical and accounting purposes. However, these machines were almost entirely designed for commercial application. This meant that mathematicians and engineers had to develop computers on their own, working independently from one another. I was no exception.

At the beginning of the 30s, while studying civil engineering in Berlin, I decided to develop and build bigger calculating machines, more suitable for engineering purposes. I approached the problem from various angles:

Firstly, from a logical and mathematical point of view:

This involved
1. program control,
2. the binary system of numbers,

3. and floating point arithmetic.

Today, these concepts are taken for granted, but at the time this was new ground for the computing industry.

Secondly, from the design angle:
1. allowing fully automatic arithmetical calculation,
2. a high-capacity memory,

3. and modules or relays operating on the yes/no principle.

My research was initially aimed at pure number calculation, but soon led on (1935/36) to new ideas about "computing" in general. Personally, I believe that was the birth of modern computer science. I recognized that computing could be seen as a general means of dealing with data and that all data could be represented through bit patterns, generally speaking.

That led to my basic hypothesis that:

"data processing starts with the bit"

At that time, of course, I didn't talk of "bits", but of "yes/no status". On the basis of this hypothesis I defined "computing" as

"the formation of new data from input according to a given set of rules"

This basic theory meant that all computing operations could be carried out by relays operating according to the dual status principle just mentioned. The most suitable devices available at the time were telephone relays.

Now a link with mathematical logic had been forged. As an engineer I had no idea of the existence of such a discipline. I developed a system of "conditional propositions" for relays - something that corresponded approximately to what is known as Boolean algebra today. My former mathematics teacher showed me that this sort of calculation was identical with the propositional calculus of mathematical logic.

From the engineering point of view, the gap between this and pure mathematical logic was bridged in order to simplify the design and programming of computing machines. At roughly the same time in England, the mathematician and logician Alan Turing was in the process of solving this problem from a different angle. He used a very simple computer as a model in order to place theoretical logic on a more formal basis. Turing's work was of major importance for the theory of computer science. However, his ideas had little influence on the practical development of computing machines.

The theories needed to be put into practice. First of all high-capacity memories had to be designed. At that time, (1935), memory consisted of single registers operating a system of numbered wheels using the decimal system. Typical problems were the input and retrieval of information, as well as the choice of counters. Capacity was fairly restricted, although some punch card machines were able to deal with up to 20 counters. These machines generally functioned on the basis that a number could be "added on".

But a new problem had to be overcome: pure memory was needed without the adding-on facility, but with high capacity and a special selection facility, as well as an elegant way of communicating with the periphery. I thought it was a good idea to base such a memory device on binary numbers from the outset. My idea was to divide the machine up into cells which would be able to hold data for a complete number, in other words, the operational sign, exponent and mantissa (where a floating point was being used), as well as additional specifications. Using the yes-no principle a "word" - as we would call it today - could be formed from a series of bits. The memory elements only needed to store yes-no values.

One device that could deal with this type of operation was the electro-magnetic relay, which can adopt two positions, "open" or "closed". However, at the time I felt that the problem could be better solved mechanically. I played around with all sorts of levers, pins, steel plates, and so on, until I finally reached what was a very useful solution, for those days. My device consisted mainly of pins and steel plates, and in principle could be extended to 1,000 words. A proper machine using telephone relays would have needed 40,000 relays and filled a whole room.

The basic principle was that a small pin could be positioned right or left of a steel lug, thus memorizing the value 0 or 1. Input and retrieval were also effected via a steel-plate construction, and the individual parts could be stacked on top of one another in a system of layers. The address system also used binary code.[2] These machines had the advantage of being made almost entirely of steel, which made them suitable for mass production.

Individual memory elements could be easily arranged in matrix form, which was very useful as far as constructing computers was concerned. Not only was a number memory now available, but it could also be used to store general data drawn from practically any source. Logic studies conducted at the same time had already shown that general calculations with any sort of data structure were possible, and that this data could be made up entirely of bit combinations. That is why I had already called the storage system a "combination memory" in the patent application.

This was something new on the Babbage designs. It was clear that programs could be stored provided they were composed of bit combinations - one reason why programmable memory had already been patented by 1936.

In the course of pursuing the basic principles of mechanical memory I developed a mechanical relay technology. This I applied to both programming and calculating parts. At the time it was not clear whether all operations could be run according to the yes-no principle, or even whether that was a good idea. That was something that was only discovered later after much hard work. Initially I developed various adding machines for binary numbers which used elements providing up to three or four positions. This was done using both electro-magnetic and mechanical relays. Finally I found a solution which worked on the yes-no principle alone. By this time the similarities between essentially very different technologies were becoming increasingly obvious. I was faced with the choice of using either telephone relays or mechanical technology for my computing machine. As mechanical memory had proved successful (I was able to build a working model in six weeks), and because of the frightening number of relays needed in the alternative system, (around 1,000), I decided in favor of the mechanical version, at first.

Inventors are often faced with that sort of decision. Today, I know that opting for relays immediately would have been better. However, working on a completely private basis, with- just the help of some friends, I started to construct a mechanical model of the computer. At first I thought it would be possible to produce it quickly. In fact it took two years to set up a half-way functioning machine which I could present to the experts. Unfortunately the surviving photos are not very good and the machine itself proved somewhat unreliable! In fact, with the help of switching algebra, it proved easy to convert mechanical relay circuits for use in electro-magnetic relay technology.

At this point I would like to mention my friend Helmut Schreyer who was working on the development of electronic relays at that time. Helmut was a high-frequency engineer, and on completing his studies (around 1936) started working at Prof. Stäblein's Institute at the Technical University in Berlin-Charlottenburg. Helmut, who was a close personal friend of mine, suddenly had the bright idea of using vacuum tubes. At first I thought it was one of his student pranks - he was always full of fun and given to fooling around. But after thinking about it we decided that his idea was definitely worth a try. Thanks to switching algebra, we had already married together mechanics and electro-magnetics - two basically different types of technology. Why, then, not with tubes? They could switch a million times faster than elements burdened with mechanical and inductive inertia.

The possibilities were staggering. But first basic circuits for the major logical operations such as conjunction, disjunction and negation had to be discovered. Tubes could not simply be connected in line like relay contacts. We agreed that Helmut should develop the circuits for these elementary operations first, while I dealt with the logical part of the circuitry. Our aim was to set up elementary circuits so that relay technology could be transferred to the tube system on a one-to-one basis. This meant the tube machine would not have to be redesigned from scratch. Schreyer solved this problem fairly quickly.

This left the way open for further development. We cautiously told some friends about the possibilities. The reaction was anything from extremely skeptical to spontaneously enthusiastic. Interestingly enough, most criticism came from Schreyer's colleagues, who worked with tubes virtually all the time. They were doubtful that an apparatus with 2,000 tubes would work reliably. This critical attitude was the result of their own experience with large transmitters which contained several hundred tubes. Apart from that, conditions were not exactly propitious for the development of a fully tube-operated machine. The War had broken out in the interim, making the procurement of staff and material very difficult. Nothing could be done by private initiative. We therefore proposed the construction of a 2,000-tube computer for special use in anti-aircraft defense to the military authorities. Although the reaction was initially sympathetic towards the project we were asked simply, "How much time do you think you need for it?". We replied, "Around two years. The response to this was, "And just how long do you think it'll take us to win the war?". The outcome was considerable obstruction and delay in the development of a German electronic computing machine. Schreyer was by now fully engaged in other projects. By the end of the War he had constructed a small experimental machine for 10 binary digits and around 100 tubes. But this machine was also lost in the general confusion just after the War.

After the War was finally over, news of the University of Pennsylvania ENIAC machine went all round the world - "18,000 tubes!". We could only shake our heads. What on earth were all the tubes for? Schreyer and I parted company after the War. At that time it was prohibited to develop electronic equipment in West Germany. As Schreyer saw no means of continuing his very interesting research he emigrated to Brazil to take up a university chair.

Schreyer died in 1985.

The English development known as COLOSSUS was unheard of outside the circle of those working on it. It was only much later that the wraps came off this very interesting project. In 1980 Schreyer and I had the opportunity to speak to the COLOSSUS people in England. We compared our circuits and it turned out that there were considerable similarities. The English had also been working on logical operations and other similar design principles.

By the end of 1938 it seemed clear that electro-magnetic relays offered the best chance of producing a reliable operating computer quickly. Before I redesigned the Z1 to operate completely with relays I made a test with a small pilot machine, the Z2. I used the mechanical memory of the Z1 with a low storage capacity (16 words), as well as the card punch and reader to build a simple computer with 200 relays operating with 16 bits and on the basis of fixed- point arithmetic.

Young transmitter specialists, including Schreyer and other friends of mine, helped me design the circuits and choose the appropriate components. But although their advice was a great help, to a certain extent I expressly set out to explore new ground. The most important thing seemed to be to keep the frequency absolutely even, so that one cycle equaled one addition, itself comprising several steps. Frequency was set using rotating disks or rollers which were covered in alternating strips of conducting and non-conducting material, contact being made via carbon brushes This principle had many advantages. Tests could be run on the machine at any speed. Another advantage was that spark extinction took place at the brushes and not at the relay contacts when circuits were being shutdown. Despite well-meaning advice from some friends, I did not make use of certain well-known telephone communication tricks such as delayed-response relays.

All in all, I was able to gather enough experience with the Z2 in order to convert the complete Z1 design for relay operation. What emerged was the Z3, which is now considered to have been the first properly functioning computer in the world. In order to make fast progress the memory was also given a 64-word capacity, making use of relays.

The Z3's basic specifications were:

  • a binary number system
  • floating point arithmetic
  • 22-bit word length, with 1 bit for the sign, 7 exponential bits and a 14-bit mantissa
  • 2,400 relays, 600 in the calculating and program section and 1,800 in the memory.

The calculations possible were addition, subtraction, multiplication, and division, taking the square root, as well as some ancillary functions. Construction of the machine was interrupted in 1939 when I was called up for military service. It was typical of the attitude prevalent in Germany at the time that I should be later released from active service, not to develop computers, but as an aircraft engineer. However, in my spare time, and with the help of friends, I was able to complete the machine. By 1941 it was working and I was able to show it to the aircraft construction authorities. The German Aircraft Research Institute in Berlin-Adlershof showed greatest interest. Professor Teichmann, who had been working on the problem of wing flutter, was particularly attracted. Unlike aircraft stress, wing flutter results in critical instability due to vibration of the wings, sometimes in conjunction with the tail unit. Complex calculations were needed in order to overcome this design problem. The most difficult part was calculating the so-called "Küssner determinants" based on complex numbers and unknown quantities in the main diagonal. I achieved a breakthrough using my equipment for this calculation. Unfortunately the Aircraft Research Institute had not been given a high enough priority for me to be released from military service. Only Professor Herbert Wagner, who was working on the development of remote-controlled flying bombs, and for whom I worked as a stress analyst, was in this enviable position. However, Wagner was very understanding, and helped as much as possible by allowing me to use some of my work time on the project. By then I had already set up my own small engineering business, the "Zuse-Ingenieur-Büro" in Berlin. The Z3 was later destroyed after bombing raids. Because of its historic importance we rebuilt it 20 years later; a replica now stands in the Deutsches Museum in Munich.

Around 1942 it was decided to build a more powerful, improved Z4. We thought that we would be able to have it ready within one, to one-and-a-half years. It was to have a mechanical memory with a capacity of 1,024 words, several card readers and punches, and various facilities to enable flexible programming (address translation, conditional branching).

Construction of the machine started well but it was not long before the War imposed its delays. In the end, construction was not completed until the close of the War. Procurement of staff and materials became increasingly difficult, and around 1943 the Berlin blitz began, with heavy bomber raids nearly every day. Several times we had to move location with the machine. During the last few weeks of the War we found refuge in Göttingen. The Z4 was the only model we were able to save, and this in the face of considerable difficulties. On the 28 April 1945 we were able to demonstrate the Z4 to Professors Prandtl, Betz and Küssner. But the Western and Eastern fronts were drawing closer daily and nobody could say whether Göttingen would be bombed or not, or whether the Z4 was safe there. The Ministry of Aviation ordered us to take the machine to the underground works in the Harz. It was there that we first learnt of the terrible conditions under which the so-called reprisal weapons - the V1 and V2 - were being built. We refused to leave the machine there, and, with the help of Wernher von Braun's staff, we managed to get hold of a truck to transport it elsewhere. And so the Z4 odyssey continued. We then moved south, ending up in a small Alpine village called Hinterstein in the Allgäu, where we were finally able to find a good place to store the machine.

Honors and Awards:

Honorarprofessor, Georg-August-Universitat, Göttingen), 1966; Honorary Degrees; Dr.-Ing.E.h., T.U. Berlin-Charlottenburg, 1956; Dr.rer.nat.h.c., University of Hamburg, 1979; Dr.rer.nat.h.c, T.U. Dresden, 1981; Dr.techn.h.c., Universitait Reykjavik, Iceland, 1986; Dr.rer.nat.h.c., University of Dortmund, 1991; Dr.h.c.sc.techn., ETH Eidgenossische Technische Hochschule, Zurich, 1991; Dr.-Ing.E.h., Hochschule f. Architektur und Ballwesen, Weimar, 1991; Dottore ad honorem in Matematica, University of Siena, Italy, 1992; Inländische Auszeichnungen/Ehrungen: Werner-von-Siemens-Ring, Stiftung Werner-von- Siemens-Ring, 1964; Dieselmedaille in Gold, DEV Deutscher Erfinder- Verband/Nurnberg, 1969; Grosses Verdienstkreuz des Verdienstordens der Bundesrepublik Deutschland, 1972,, mit Stern, 1985); Ehrenmitglied der Deutschen Akademie der Naturforscher LEOPOLDINA, Halle/Saale, 1972; Aachener und Munchener Preis, Carl-Arthur-Pastor- Stiftung, Kuratorium der Aachener und Munchener Versicherungs AG, 1980; Ehrenplakette der Stadt Bad Hersfeld, 1980; Konrad-Zuse-Medaille, ZDB/Zentralverband des Deut- schen Baugewerbes) 1983 Bernhard-Weiss-Plakette, VDMA/Verband Deutscher Maschinen- und Anlagenbau e.V./Dusseldorf, 1981; Bayerischer Maximiliansorden, Bayerischer Ministerprasident, 1984; Goldener Ehrenring, Deutsches Museum/Munchen, 1984; Cothenius-Medaille, LEOPOLDINA, Deutsche Akade- mie der Naturforscher/Halle - Saale, 1985 ; Ernst-Reuter-Plakette, Senat Berlin, 1985; VDE-Ehrenring, Verband Deutscher Elektrotechniker e.V./Dusseldorf, 1986; Philip-Morris-Ehrenpreis, Philip Morris GmbH/DABEI, 1987; Wilhelm-Leuschner-Medaille, Hessischer Ministerprasident, 1987; Ehrenmitglied: Deutsche Akademie der Naturforscher LEopollDINA, Halle/Saale, 1972; Akademischer Verein Motiv, 1982; Verein des Schleswig-Holsteinisches Museums fur Rechen- und Schreibtechnik e.V., Altzenholz, heute: MICOM, 1983; Gesellschaft fur Informatik e.V./GI 1986 Verein islandischer Ingenieure, Reykjavik/Island, 1985; Deutsches Museum, Munchen, 1990; Vereinigung der Freunde und Förderer der Ingenieur- schule an der Fachhochschule Schmalkalden e.V., 1992; Ehrenburaerrecht: Ehrenburgerrecht der Stadt Hunfeld, 1975; Namensqebunq: Konrad-Zuse-Strasse, in Bad Hersfeld/Hessen, 1972; Konrad-Zuse-Schule, in Hünfeld/Hessen, 1978; Konrad-Zuse-Medaille, ZDB Zentralverband des Deutschen Baugewerbes und GI/Gesellschaft f. lnformatik e.V., 1981; Konrad-Zuse-Zentrum für Informationstechnik Berlin/ZIB, in Berlin, 1984; Konrad-Zuse-Zertifikat, Freundeskreis der Berufl. Schulen e.V./Bad Hersfeld, 1985; Zuse-Raum, Berufliche Schulen/Bad Hersfeld, 1985; Konrad-Zuse-Gesellschaft, Grundung am 6. 9. 88 in Hünfeld, 1988; Konrad-Zuse-Haus, Fa. PDS Programm + Software GmbH, Rotenburg/Wumme, 1989; Konrad-Zuse-Programm, Förderung v. Gastdozenten ausländischer Hochschullehrer - DAAD/Deutscher Aka-demischer Austauschdienst, Bonn, 1991; Konrad-Zuse-Zimmer, Schelztor-Gymnasium, Esslingen, 1991; Zusestrasse, in 0-Hoyerswerda, 1991.

Copyright J. A. N. Lee, September 1994.

Last updated 94/09/30