On the Beginnings of Computer Development in Poland

Leon Łukaszewicz

(Appeared in Annals of the History of Computing in 1990.)

Abstract: Development of computers in Poland began at the end of 1948 with the formation of the Group for Mathematical Apparatus of the Mathematical Institute in Warsaw. The beginning was not easy because at that time Warsaw was still rebuilding after the destruction of World War II. The first analogue computer called the "Analyzer of Differential Equations", was completed in 1954 and then regularly used for several years. The first successful digital computer called XYZ was completed in 1958. It performed about 800 operations per second and became a milestone in the development of Polish computers. Soon the XYZ computer was improved and, under the name ZAM 2, was manufactured and installed in many places in Poland and abroad. An asset of XYZ and ZAM 2 was the System of Automatic Coding, introduced in 1960 and often called the Polish FORTRAN.

Forty years have already gone by since seemingly ordinary events in which I participated initiated the computer development in this country. As a result of those events, the Group for Mathematical Apparatus, GAM (Grupa Aparatow Matematycznych), of the Mathematical Institute in Warsaw was constituted in December 1948. The history of this group has already been described in detail (see Marczyński 1980, and references in Łukaszewicz 1989) and so I will only present the most important facts and try to evoke the atmosphere of those memorable and lovely days.

For me everything began in the following way. Having just received my M.Sc. in radio engineering, I started my first job in the Institute of Telecommunication in Warsaw and simultaneously continued studying mathematics at the University of Warsaw. For this reason a number of my colleagues at the Institute used to ask me to solve various mathematical problems and do the relevant computations. In doing this I used only a pencil, paper, and a slide rule. Professor Janusz Groszkowski, the Director of the Institute, often asked my assistance when working on his theory of nonlinear oscillations. During one session Professor Groszkowski told me about the construction of an electronic computer which was then being planned at the Institute of Mathematics in Warsaw. He advised me to contact Professor Kuratowski, who was organizing the Institute, if I was interested. I was thrilled with the news, all the more since I had read about the ENIAC in Electronics. I was impressed by the design and calculating potential of this machine, in particular when I realized that it was able to calculate in seconds what had often taken me an entire day's work.

I eagerly approached Professor Kuratowski, whom I visited in November 1948. He told me that during his recent stay in the U.S. he had heard about the possible great advantages of electronic computers for applied mathematics. According to Kuratowski, in the U.S. they were planning the construction of a dozen or so such computers and he was convinced that there should be at least one in Poland. Accordingly he wanted to form a special group at his Institute and he would gladly allow me to participate. Dr. Henryk Greniewski was to be the head of this group, but at that time there were no other members. When asked about my suggestions, I named my two friends, young engineers, Krystyn Bochenek and Romuald Marczyński, who eagerly accepted the offer. Thus, in December 1948 it was decided to form the GAM with this staff.

It was quite clear then that our original goal, with the ENIAC as a model, was unrealistic because the machine was a giant equipped with more than 18,000 electronic tubes. The ENIAC was one of the top achievements of advanced American technology of that day, but for all that, it did not function smoothly - a fact we were not aware of. On the other hand, analogue computers, being in competition with digital ones, demanded operational precision which could not be achieved without high-quality components. But we had neither proper components nor sufficient experienee in constructing complex and reliable electronic equipment. Thus it was only the continuing progress in computer technology and our youthful enthusiasm that gave us a chance of achieving our aim. The management of the Institute realized the situation, but there was little choice but to trust us, watch our efforts, and wait patiently for the outcome.

For 18 months GAM did not even have any premises of its own. It should be remembered that we were in Warsaw, a city almost completely demolished by the German army in 1944. For us it was a period of planning our future laboratory, studying books and periodicals, which were reaching us gradually, and discussing theoretical problems in seminars, e.g., a mathematical description of a computer. Dr. Greniewski, a logician and philosopher, was in charge of the seminar. He was an excellent head of our group and although he could not contribute much to the design of computers he had considerable general experience, which he offered us freely.

Finally, in the autumn of 1950, the Institute of Mathematics was given its own premises. It was part of the reconstructed building which had been the seat of the former Warsaw Scientifc Society. It was there that our group was allotted the stupendous number of three separate rooms. In the first we put desks, the second was a kind of storeroom for various electronic and other components, and the third one, the largest, was our lab. In one corner of it, Krystyn worked on the construction of his analogue Analyzer of Linear Algebraic Equations, ARAL (Analizator Równań Algebraicznych Liniowych); in the second corner I was busy with my analogue Analyzer of Differential Equations, ARR (Analizator Równań Różniczkowych); and in the remaining two corners Romuald was dealing with his digital Electronic Automatically Computing Machine, EMAL (Elektroniczna Maszyna Automatycznie Licząca).

Three years later our space extended and a large group of young and talented computer enthusiasts joined us. I will name only a few of them: electronic engineers Zygmunt Sawicki, ZdzisŁaw Pawlak, Jerzy Fiett, Wojciech Jaworski, and Jerzy Dańda; mathematicians Antoni Mazurkiewicz, Tomasz Pietrzykowski, Józef Winkowski, Jerzy Swianiewicz, and Krzysztof Moszyński.

I loved my job at the Institute not only because it was very interesting; it was also an honor to be working together with so many famous Polish mathematicians, such as Professors Kazimierz Kuratowski, Director of the Institute of long standing, Stanisław Mazur, Hugo Steinhaus, and Stanisław Turski, who offered us their kind assistance whenever we were in trouble. Professor Wacław Sierpiński often discussed with me the possibilities of computing very large prime numbers; and with Professor Oscar Lange, the economist, we talked about the possibilities of modeling our national economy on computers. It was very exciting for us that such outstanding scientists showed their interest in our efforts and were looking forward to the results.

Our job, although very stimulating, was poorly paid. It was quite easy to obtain a salary twice as high in industry and also be alloted a flat, which in those days of the acute housing problem mattered a great deal. We turned to Professor Kuratowski with our grievances. He answered that "low salaries are a severe but necessary test of a genuine dedication to science of young research workers. If the salaries were high, what sort of people would we get here?" His arguments were not entirely convincing for us, but of course no one quit the Institute we adored. It should also be added that, some time later, some of us, already "tested", were allotted the flats we and our families had dreamt about.

The first important achievement of GAM was the ARR, which began operation in 1954. Its size was quite impressive as it contained about 400 electronic tubes. It made it possible to solve systems up to eight ordinary, but not necessarily linear, differential equations. The parameters of the equations were changed easily by setting knobs and their solutions were immediately seen on several screens, which could not be done by a digital computer in those days. ARR was applied to a great many problems, such as the investigation of nonlinear mechanical vibrations. It was the first Polish computer to be regularly used and it attracted several talented mathematicians to the Institute.

The recognition of ARR was confirmed by the fact that its designers were awarded the State Prize in 1955. It arrived at a proper time since the first effects of our research done in GAM had been slow to make their appearance, and the patience of the Institute and the Academy supervising it had been wearing out. Thanks to the prize we were trusted again; in this way ARR prepared the way for our further computers.

An essential cause of the delay was the lack of sufficiently precise and reliable components, in particular electronic tubes. Initially we used components of Polish make, but their quality was unsatisfactory. We also used components that the German army had left behind, but their application was rather limited. It was only much later that we were able to import some components of higher quality from the West.

Owing to the lack of proper components it was not possible to make EMAL work. That is why at the beginning of 1956 the Institute's management decided to unite the whole staff ofGAM, with me as head, in the task of the design and construction of another digital computer. This time our attempt was successful. Within three years the first Polish operating digital computer, called XYZ, was designed and constructed. Having an ultrasonic memory, it performed about 800 operations per second and for several years remained the fastest digital computer built in this country. The presentation of XYZ in the autumn of 1958 to our management and the general public aroused great interest.

While designing XYZ, we were aware of the limitation of our experience and that is why we made use, whenever possible, of foreign concepts. The architecture of XYZ was a simplification of the already fairly simple architecture of the IBM 701 (Buchholz 1953); we believed that such a renowned company as IBM could not make mistakes. On the other hand, the design of elementary electronic cells (flip-flops) of XYZ was taken from the Soviet BESM 6. They were "dynamic" cells whose description and then demonstration we were offered in Moscow in 1956. But the ultrasonic memory of XYZ, based on mercury tubes, was taken from EMAL. It was designed by Romuald Marczyński with the assistance of Henryk Furman and then improved upon by Zygmunt Sawicki and Jerzy Dańda. That memory, however was never sufficiently reliable and was replaced in later constructions by ultrasonic memory based on nickel wires, developed by Zofia Siwak. The XYZ computer was also equipped with a magnetic drum memory. The process of construction of XYZ was efficiently managed by Zygmunt Sawicki.

Shortly after being completed, XYZ began to be operated regularly, servicing the Bureau of Computation and Programming, BOP (Biuro Obliczeń i Programów), which emerged from our group.

In general, the XYZ computer was a milestone in the development of the Polish computers. It proved that the construction of reliable and efficient digital computers was definitely possible in this country. In effect, many other research centers became interested in computers and the Government's attention was drawn to their possible production and wide application. Thus XYZ aroused country-wide interest in computers in Poland.

Soon the XYZ computer was improved and its new version, named ZAM 2, was manufactured and installed in many places both at home and abroad. (In the meantime GAM was renamed the Laboratory of Mathematical Apparatus, ZAM (Zakład Aparatów Matematycznych), and then the Institute of Mathematical Machines, IMM (Instytut Maszyn Matematycznych), of the Polish Academy of Sciences.) In 1964 the team of designers of XYZ and ZAM 2 were awarded the State Prize.

One of the assets of XYZ and ZAM 2 was the software, in particular the System of Automatic Coding, SAKO (System Automatycznego KOdowania), (Łukaszewicz 1961, Mazurkiewicz 1961) introduced in 1960 and often called the Polish FORTRAN. W.M. Glushkow and S.S. Sobolew, the Soviet academicians who attended a software conference in Warsaw, 1961, stated that they had not yet met such a system in the U.S.S.R. A similar opinion was expressed by M. Kieldysh, President of the Soviet Academy of Sciences, during his visit at our Institute in 1964. He was invited to formulate a simple but nontrivial numerical problem; he proposed a numerical solution of a differential equation which he had formulated. The problem was promptly coded in SAKO by Antoni Mazurkiewicz and in a quarter of an hour correct results were printed by ZAM 2 in the form of a well-described table. Kieldysh was surprised by the speed of the whole operation.

In this way in Poland, at the end of the 1950s and the beginning of the 1960s, fairly fast progress of computer development and application was made. In the class of small computers, ZAM 2 were similar to many computers produced in Western Europe, the Soviet Union, and Japan at that time, and in software our position among the socialist countries was strong. A factory producing computers was founded in Wroclaw and its engineering staff was trained at the Institute. Several Polish universities organized computer science departments and many other institutions started computer application (Marczyński 1980). But what finally ensued is another story.

References:

Professor Łukaszewicz was born in 1923 in Warsaw. He graduated in electronics in 1948, in mathematics in 1950, and received his PhD from the Warsaw Institute of Technology in 1952. In 1948 he joined the Group for the Mathematical Apparatus of the Mathematical Institute of Warsaw. He pioneered the development of computers and computing in Poland, designed the first Polish computers, both analog and digital ones, and established the Institute of Mathematical Machines, which he directed from 1956 to 1966. He also designed a number of programming languages such as SAKO and EOL, which have been widely used. Recent research includes definition of programming languages and specification of programs. He also been visiting lecturer to many U.S. universities.
Professor Łukaszewicz was the representative of the Polish Academy of Sciences to IFIP from 1961 to 1985 and Vice President of IFIP from 1964 to 1968. He has been a member of the Academy from 1976.