Beginning of Computing in the Soviet Baltic Region
This talk includes references at events, people and trends in computing in the Baltic region of the Soviet Union. It is based on the material presented at the Conference on History of Nordic Computing held in Turku, Finland in 2007.
Title of the work sets the time period of the subject - until fall of the Soviet Union, practically in the end of the 1980s. Most of the material of the present paper is available from a talk given at the Conference on History of Nordic Computing held in Turku, Finland in 2007 . The geographical focus is on three Baltic republics: Latvia, Lithuania, Estonia and Leningrad. My personal experiences relate to Estonian computer engineering and science. Therefore, we will consider computing in this small country in more detail.
Computing in the Soviet Union was considered as a part of cybernetics - a “capitalist pseudoscience” in fifties of the last century; hence, it was publicly nonexistent although the computers were used by physicists and space engineers. Nikita Khruschov suddenly decided in 1958 that the country urgently needed a larger number of computer engineers and mathematicians with computing skills for defense and space industry. Several hundreds of young physicists, mathematicians and electronic engineers were reeducated in two years in Leningrad Polytechnical Institute (LPI) and Moscow Energy Institute (MEI) providing them as good education in computing as it was possible in those days. It may be interesting to look at the computer science curriculum of those days. Below is the complete list of courses together with number of hours of supervised work - lectures and applications (taken from the course list of the author) given to the computer specialists in Leningrad.
Ordinary differential equations – 90
Algebra – 90
Functions of a complex variable – 90
Probability theory and statistics – 105
Partial differential equations – 30
Numeric methods – 60
Programming – 55
Control theory and tracking systems – 204
Semiconductors and magnetic elements – 60
Electronic devices – 150
Theory of electric circuits – 60
Arithmetic and logic of computers – 45
Theory and design of analog computers - 90+180
Theory and design of digital computers - 108+206
The last two courses included many hours of practical work (180 and 206 respectively). One can notice the absence of logic, theory of algorithms and discrete mathematics, although the curriculum was rather mathematically oriented. As we can notice, only a single rather short course in programming was offered. The education was strongly oriented at hardware design and applied mathematics, because the educators knew little about programming. The term “software” did not exist yet. Although programming languages Fortran, COBOL and Algol already existed, they were neither used nor taught to students in those days in the Soviet Union. However, the first book about a compiler was already written by Andrei Ershov , and the first two Algol compilers were under development under supervision of Andrei Ershov and Svjatoslav Lavrov.)
2. First Years of Baltic Computing
When I went to study computing in Leningrad in 1959, there was computing experience in a number of institutes. Computers were designed at least in the Leningrad Polytechnic Institute (team leader - professor Taras Nikolajevich Sokolov), in the Leningrad Electrotechical Institute (Vladimir Borisovich Smolov) and in a closed military institute KB-2 of Electronic Technology led by Philip Staros (see also below). Computer factories were in the military sphere, and were almost inaccessible even for the computing students. Programming was taught to some extent in the universities, and the programming expertise existed in computing centers, e.g. in the Computing Center of the Leningrad Department of the Mathematical Institute of the Academy of Sciences. This expertise gradually spread to the whole Baltic region.
Among the graduates of the classes given in Moscow and Leningrad were ten Estonians and even a larger number of Lithuanians who then returned to their countries. This was an essential source of expertise in computing in the Soviet Baltic Republics.
The first input to Estonian computing came from some enthusiastic mathematicians of Tartu University. At the end of the 1950s, Ulo Kaasik initiated mathematically oriented computer science education at the University of Tartu, and its first graduates came in 1960. A young mathematician Leo Vohandu soon joined Ulo Kaasik in Tartu, and moved later to Tallinn Technical University where he coordinated the computing education.
Lithuanian computing graduates from the Leningrad Technical University became the key players in a newly founded computer plant in Vilnius, and therefore they were not immediately visible in science. A computer plant in Vilnius, later known under the name “Sigma”, became one of the major computing equipment producers for nonmilitary computer systems. The Ruta 110 computer designed and produced in “Sigma” was widely used in the Soviet Union. There were two Lithuanian centers of computer science research - one in Vilnius and another in Kaunas. The leader of the center in Kaunas became Henrikas Pranevicius who graduated Kaunas Polytechnic Institute as a radio engineer in 1964. People know him for his works in formal methods and simulation applied to distributed systems. Albertas Caplinskas and Olegas Vasilecas worked in the field of knowledge-based software in Vilnius.
A Latvian young mathematician Janis Bardins from Riga was a graduate student of Boris Trakhtenbrot in Novosibirsk, a well-known expert in automata theory in sixties. Barzdins became the leader of computer science in Latvia. Janis Barzdins obtained fundamental results in inductive inference, and later applied his experiences in inductive program synthesis. Besides these works, they did more research on the border of logic and computing at the University of Latvia. An active group of researchers in computer science grew around Janis Barzdins, including I. Etmane,
R. Freivalds and others. Their research focused on logic and included various methods of synthesis of programs. Another research direction in the University of Latvia was automatic test generation (Janis Bicevskis, Audris Kalnins, Juris Borzovs). Interesting research was carried out in the Riga Technical University in the field of fuzzy sets (Janis Osis had spent a year with L. Zadeh at Berkeley University) and system analysis by means of topological models.
Early Estonian computing was influenced by the fact that Institute of Cybernetics was founded in Tallinn in 1960, primarily by initiative of Nikolai Alumae, who needed computers for his research in dynamics of thin shells (submarine hulls). This institute was a place where the first digital computer M-3 was built in Estonia in 1960. It had been originally designed in Minsk, but was significantly improved by adding a core memory instead of a much slower magnetic drum. This increased the performance of the computer considerably. The Institute of Cybernetics became a leading research center in computer science and computer applications in the Soviet Baltic region.
The first minicomputer called STEM, see Figure 1, was designed and built very early in Estonia - in years 1962 - 1964 at the Electrotechnical Research Institute in Tallinn. It had 16-bit words, small core memory, large ROM that hosted a kind of a database of metal cutting parameters and operations’ times. It supported an interactive textual input-output through electric typewriter. This computer was unusually reliable for those days. (Its logic was built on reliable ferrite-diode components produced by a military factory in Leningrad.) It was used in a technology department of Kirov factory in Leningrad and not in a computing center, because it did not require technical support around the clock that was otherwise a common requirement in those days. It was used for engineering calculations during many years. More computers of this kind, but on different component base were built in Tallinn later for large aviation industry plants.
a) Typewriter as 10 device
b) ALU on ferrite-diode elements took most of the space
Fig. 1. Minicomputer STEM in 1964
Neither M-3 Estonia nor STEM were the first computers built in Estonia. The very first was an analog computer designed and built in 1959 by electrical engineers of the Tallinn Technical University for the purpose of modeling and simulation of large power networks. This computer is shown in Figure 2.
3. Computer and Software Science in Estonia
Early research in computer science in Estonia was mainly in programming languages. This was language design, compiler development and theory of formal languages. Malle Kotli developed and implemented a language called MALGOL (modular Algol) that was widely used on popular Minsk computers. A rather original data processing language VELGOL was developed and implemented by a team led by Velio Kuusik. On the theory side, Mati Tombak became a leader of research in formal languages, and he supervised a number of Ph.D. theses in this area. Success in syntactic approach inhibited deeper interest in semantics of computations in Estonia for years. The situation changed in the end of seventies, when Merik Meriste and Jaan Penjam proposed new efficient methods of implementation of attribute semantics, and wider interest in automatic program construction emerged.
3.1. Theoretical Computer Science
In the beginning of seventies, Wilhelm Kracht introduced automata theory in his seminars to young scientists. It gave output in the form of Ph D theses on decomposition and other problems of automata (Gabriel Jakobson, Andres Keevallik, Paul Leis) soon. This research domain became practically widely recognized many years later, when fast computers and new challenges in chip design appeared. Many theses were written in the Institute of Cybernetics on the border between computer science, numeric methods and statistics. Leaders from the math side were Ivar Petersen and Sulev Ulm. A brief survey of research topics in 1970s and 1980s in Estonia is as follows.
Research in databases (Ain Isotamm, Anne Villems, Enn Tyugu, Ahto Kalja, Hele-Mai Haav).
Control theory and computer control (Ulle Kotta, Raul Tavast, Leo Motus).
Systolic algorithms and FFT (Ilmar Arro, Toomas Plaks).
Synthesis of programs (Enn Tougu, Grigori Mints)
Attribute grammars (Jaan Penjam, Merik Meriste)
Expert systems and knowledge representation (Jaak Tepandi, Enn Tougu, Mare Koit).
Logic - proof theory, realizability, model checking (Grigori Mints, Tanel Tammet, Sergei Tupailo).
Software environments (Boris Tamm, Juhan Prauden, Mihail Matskin, Aleksander Shmundak).
Test generation (Raimund Ubar).
3.2. Software Tools and Applications
On the software side, the first remarkable result was development of a language and environment SAP-2 for numeric control of machine tools in the beginning of sixties. The system SAP-2 was introduced in the Soviet aviation industry and gave a good position for its leading developer Boris Tamm in the Soviet computing. Another group of researchers (Enn Tyugu, Kalju Tinn et al) developed and applied in industry a modular programming environment SMP in the end of sixties. It included even a simple operating system for batch processing of jobs on Minsk 22 computer. This became a starting point for research in software engineering here, because SMP was supported by a well-defined software technology and documentation. This research direction was continued by development of structural synthesis of programs and its implementation in several software tools (PRIZ, MicroPRIZ, ExpertPriz, Nut and Nuts). These tools were used in the development of CAD/CAM applications. The first engineering applications were optimization programs for machine tools - calculating cutting conditions, processing time etc. for Kirov plant in Leningrad in sixties. Larger apphcations were developed for Elektrosila plant in Leningrad and rocket engines plant in Dnepropetrovsk in seventies. Numerous applications in power semiconductor design and technology for Tallinn Electrotechnical Plant were developed in eighties under supervision of Valeri Grigorenko.
On the data processing side, there was an information system project for a large wholesale warehouse of Estonian Consumers Cooperative Society (ETKVL) that was completed already in sixties. The ETKVL administration strongly supported this project, and it attracted good software developers due to good working conditions and salaries; it became a success case of large information system development in the Soviet Union. Another long-lasting and quite successful information technology project in Estonia was computer control of chemical processes in the oil shale chemistry, done by researchers of the Institute of Cybernetics and supervised by Raul Tavast in 1970s.
3.3. Computer Design Office
An important milestone of computing in Estonia was founding of the Computer Design Office (EKTA) of the Institute of Cybernetics in 1976. This had been a dream of Harry Tani, an outstanding computer engineer, who became the director of research of EKTA. Due to his personal contacts with German engineers as well as with researchers in the Soviet Union, EKTA got advanced microprocessors and printed circuits technology, and it evolved into a leading center in design and application of microprocessor systems in the Soviet Union. The Computer Design office even designed and manufactured a small number of personal computers Juku for Estonian schools in 1988. It was a dream that success of this project would have the influence on education in Estonia comparable to the publication of bible in the native language that had happened in the eighteenth century. Unfortunately, manufacturing of Juku computers in larger numbers was impossible because of shortage of reliable components and devices like disk drives.
4. Computing in Leningrad
Leningrad had strong computer science and engineering education in many universities. However, paradoxically, few widely known results in computer science came out. Probably, the main reason was the confidentiality of many works performed in military institutions of this city. This kind of institution was, for example, a Construction Bureau-2 (KB-2) of Electronic Technology headed by Philip Staros (Alfred Saiant), who developed the lightweight computers for space, and was the first to develop pocket calculators in the Eastern Block in seventies. Still, we remember a number of very interesting and pleasant people from Leningrad closely related to Estonia. Viktor Varshawski and his colleagues were most supportive to young Estonian researchers. A prominent computer scientist Svjatoslav Sergejevich Lavrov supported Estonian researchers after he had moved from Moscow to Leningrad in the beginning of seventies. He became a professor at the Leningrad University and director of the Institute of Theoretical Astronomy where he organised a strong group of researchers. New generation of researchers grew at the Leningard University, among them Andrey Terekhov. Leningrad Research Computer Center (LRCC) of the USSR Academy of Sciences was founded in 1978, and this institute grew in a leading research centre (SPIIRAN today). Estonian researchers had lasting good contacts with Viktor Vasiljevich Aleksandrov from this institute. A special relation was between the Institute of Cybernetics in Tallinn and Leningrad Division of Mathematical Institute of the Soviet Academy of Sciences (LOMI). A strict constructivist logician Nikolai Shanin and his group had a strong influence on Estonian theoretical computer science. The members of this group Sergei Maslov (proof theory), Anatol Slisenko (recursion theory), Grigori Mints (proof theory), Yuri Matijasevich (algebra and logic) have all strongly influenced Estonian computer scientists. Grigori Mints worked as a researcher in the Institute of Cybernetics for ten years before taking a position of professor at Stanford University in the USA.
4. Two Leagues of the Soviet Computing
Speaking about the computer science in the context of the Soviet Union, one has to bear in mind that it had been from the very beginning closely related to the Soviet power structures (defense industry and military institutions). This continued even later when usage of data processing became widely available. In the conditions of shortage of resources, the computing industry and computer science could be divided roughly into league A that had better resources (including practically unlimited number of people in the research groups) and served the power structures, and league В that had shortage of resources, but more openness and some freedom of research. Attributes of league A were computers M-20, BESM-4, BESM-6 and later Elbms. League В had mainly the popular Minsk computers, and from the end of the seventies, also ES (or Rjad) computers that were copied from the IBM 360/370 mainframes. Computers mattered for software research in those days, because software was very much dependent on a hardware platform.
Estonian computer science belonged to the league B, especially, there was no BESM computer in Estonia except in some military institutions that were completely closed to researchers. It seems now that this was a smart decision of leaders of local institutes, in particular, of Boris Tamm and Nikolai Alumae who were in the position of influencing the decision-making on high level. This gave more freedom in communication with West and more openness. However, already from the beginning of seventies, Estonian researchers established good contacts with researchers from the league A. This happened due to regular summer and winter schools organized in summer and winter resorts of Estonia jointly by universities and Academy of Sciences with good programs and participation of research leaders of most of the league A groups. Gradually, we started feeling like belonging to the league A, except that we did not have the right computers. Finally, a decision was made at the end of seventies that we should try to obtain a new Soviet supercomputer Elbrus-1. It happened that Elbrus-1 with serial number 10 was planned for delivery to the Institute of Cybernetics in 1979-1980. The academy also got financing - almost 11 million rubles. Only the computer did not appear in time. It did not appear even a year later, and became operational only in 1987. Then we had already workstations that made simple arithmetic operations with short numbers faster than Elbrus-1 with its long words and very complex CPU.
5. The Start Project
START was a large computer hardware and software project in the Soviet Union where Estonian researchers actively participated in eighties. This project was initiated by researchers from Moscow (Viktor Brjabrin), Novosibirsk (Vadim Kotov and Aleksandr Narinyani) and Tallinn (Enn Tyugu) as a late response to the Japanese Fifth Generation Computer Project. There had been some attempts to establish a cooperative fifth generation computer project in the Eastern Block as a response to the respective Japanese project. These attempts were not successful, and this gave an opportunity for the researchers with good contacts to Guri Marchuk (the Chairman of the State Committee for Science and Technology and President of the Academy of Sciences) to propose the project START. Its intent was a completely open (non-secret) virtual research enterprise, contrary to several other similar Soviet projects. The aim of the initiators of the project was obviously to get better resources for testing their ideas in computer architecture (Vadim Kotov), software (Viktor Brjabrin) and artificial intelligence (Aleksandr Narinyani, Enn Tyugu). Due to the direct support from the president of the Soviet Academy of Sciences, the project enjoyed good resources. A special issue of the Communications of the ACM  described the outcome of the project START. About thirty persons participated in the project from Tallinn. The main results in Tallinn were a workstation PIRS with a 32-bit processor KRONOS (processor developed in Novosibirsk), including software of the workstation (C compiler, UNIX installation, and a windowing system) and an intelligent programming environment Nut written in C. Later on, the Nut system appeared on many workstations and PCs; it was used for simulation in large projects such as hydraulic systems design and a radar system design of Estonia.
6. Western Contacts
Although Estonia was behind the iron curtain, the country had better scientific contacts with the West than most parts of the Soviet Union. There was a special agreement on scientific cooperation between the Soviet Union and Finland in the field of computer science. Academician A. A. Dorodnitsyn supervised this cooperation, but Estonian scientists enjoyed the Finnish contacts without much interference from Moscow. In the beginning of 1988, the cooperation contract became an Estonian-Finnish agreement. Finnish scientists (Reino Kurki-Suonio, Markku Syijanen, Hannu Jaakkola, Timo Jarvi, Esko Ukonen, Kari Eloranta and many others) were frequent visitors to Estonia. Jaak Henno received a postdoc position with A. Salomaa’s group in 1976; he worked on the complexity of multiplace functions and even published together with Salomaa. Good contacts were established with Denmark (Dines Bjomer) and Sweden (Bengt Nordstrom’s group and Jan Smith in particular, also Per Martin-Lof) in eighties. Eric Sandevall from Linkoping sent a source code of Interlisp to the Institute of Cybernetics when it became a popular AI programming tool, and helped in this way the researchers in artificial intelligence to become a part of international AI community. Estonia became a meeting place of western and eastern computer scientists where they held numerous meetings, because it was easily accessible from both sides, especially by ferry from Helsinki.
7. Restructuring of the Research
The START project gave some resources to researchers in the end of eighties, when the international embargo on hardware and software was very restrictive. In addition, EKTA had good economy, and had contacts with partners from West. This helped the research in computer science to survive until the end of the Soviet Union, when the situation changed abruptly. This is visible from the Figure 3 where one sees almost constant growth of the number of employees of Institute of Cybernetics and EKTA until the year 1992 . The only disruption of the Unear growth is due to building up EKTA in 1976-1978.
What happened later is another story. Briefly, the banks and other rapidly developing enterprises attracted smart experts, and a number of researchers with good credentials left to other countries. EKTA became a small independent high-tech company, also some other application-oriented groups of the institute left it. The institute was incorporated in the Tallinn University of Technology. Today, it continues with almost constant number of employees, including some very bright young researchers, as a typical Western research center.
The story of Institute of Cybernetics and EKTA was rather typical to research centers that lost most of their financing that formerly came from the academy of sciences. Some of them disappeared, some were restructured. The bright computer experts found new jobs easily.
Fig. 3. Number of employees in the Institute of Cybernetics
A.P. Ershov. Programmiruyushchaya programma dlya bystrodeistvyyushchei elektronnoi schetnoi mashiny, 1958. (Programming program for fast-acting electronically computing machine).
Communications of the ACM, v. 34, No.6 (1991) p. 46-59.
Institute of Cybernetics in changing times. (In Estonian: Ktibemeetika Instituut muutuvas ajas) Institute of Cybernetics, Tallinn, ISBN 9985-894-25-1(2000).
1. J. Impagliazzo, T. Jarvi, P. Paju (Eds.) History of Nordic Computing 2 — Second IFIP WG 9.7 Conference, HiNC2, Revised Selected Papers. IFIP Advances in Information and Communication Technology, v. 303. Springer, 2009.
Об авторе: Institute of Cybernetics at Tallinn University of Technology
Материалы международной конференции Sorucom 2014 (13-17 октября 2014)
Помещена в музей с разрешения авторов 13 апреля 2016