During the preparation and defense of my doctoral dissertation at Moscow State University, I lived with several graduate students from Ukraine who introduced me to the academician Boris Vladimirovich Gnedenko, from the Ukrainian Academy of Sciences. At that time, Gnedenko was working as the director of the Institute of Mathematics and the academic secretary of the Department of Mathematics and Mechanics. In March 1956, he invited me to come to Kiev; it was my first trip there. After a brief tour of Kiev University, Gnedenko showed me personnel files of young specialists who were about to graduate and had been selected for work at the Institute of Mathematics in Lebedev's former laboratory. This laboratory had been moved from the Institute of Electrical Engineering. In autumn of 1956, during my second trip, my transfer to Kiev was finalized. I became the director of the laboratory of computing equipment at the Institute of Mathematics. The laboratory was supposed to be reorganized as a computing center for the Ukrainian Academy of Sciences, according to a 1955 decree to create computing centers in the academies of the Soviet Republics.

Rabinovich remembered:

It took Glushkov half a year to make a difficult decision—instead of topological algebra, where he had already obtained brilliant results, he shifted to cybernetics, which back then was frequently criticized by the authorities. The original laboratory staff had been selected by Lebedev and by now was a seasoned, cohesive group of scientists.It could have been the main reason why they were slow to welcome the mathematical theoretician Glushkov, although personally, he was liked by everybody from the very beginning. Glushkov's unquestionable talent, brilliant intellect, charm, his infectious enthusiasm about the new science, and personal diligence were instrumental in gaining the support of the laboratory staff and his progress in cybernetics. It's worth mentioning some of the major research work carried out in the laboratory employed the MESM. Specifically, the research into the theory of programming, led to Vladimir Korolyuk and Ekaterina Yushenko's creation of address mode language, followed by the creation of statistical and optimization problem solving methods by Gnedenko, Vladimir Semyenovich Mikhalevich, and others. Dashevsky was the head of this group, assisted by Solomon B. Pogrebinsky, Alla Leonidovna Gladish, and others. The same staff took part in other design work and the testing of new logic elements, in particular ferrite-diodes and semiconductors.

At the same time, the computing machine SESM had already been tested and was operating at the laboratory. This was the first computer in the Soviet Union that worked on the basis of matrix-vector processor with pipeline organization of calculations and combined input of data and output of calculation results. The architecture of the SESM represented Lebedev's ideas. But Glushkov did not dismiss it. On the contrary, he took an active interest in the project. Overcoming the designers' unwillingness to summarize a paper on the SESM's design philosophy (the computer was already completed) Glushkov insisted on writing a book about it. This was reasonable as the SESM had a number of structural novelties that could be used as independent units, such as dynamic registers on magnetic drums, a system of internal diagnostics, etc. The book was even published in the United States, apparently one of the first Soviet books published abroad.

An exceptionally important project at the laboratory was the Kiev computer. The project was established and supervised by Gnedenko; Dashevsky was responsible for the design. The machine was intended for use in the new Computing Center established at Lebedev's former laboratory and was supposed to represent a new breed of technology. It was also supposed to have these additional components: asynchronous control (the first in the Soviet Union), ferrite Random Access Memory (RAM), external memory on magnetic drums, a decimal input/output system (as in the SESM), passive memory with a set of constants and subprograms of elementary functions, and an enhanced operational system, including group operations with address modification, enabling operations to be done on complex data structures. The same group of scientists that created the MESM initiated the Kiev's design. Scientists from Institute of Mathematics – Korolyuk, Pogrebinsky, and Yushenko – worked on selecting operations. Glushkov joined this effort at the final stage of technical design, assembly, and computer adjustment, and shared leadership of the project with Dashevsky and Yushenko.

Another major project that began before Glushkov's arrival at the laboratory was the design of a double computer system for radar air target tracking and directing jet fighters to intercept them. Malinovsky and Rabinovich headed two small groups for this project. Malinovsky was in charge of the radar information processing and Rabinovich led the computer targeting design. With Glushkov's arrival, the work took on a new shape. He put everything on a strict scientific basis and formulated a mathematical theory for the targeting process. The results of his work were approved and used to create the Soviet Union's standard anti-aircraft defense system.

How did Glushkov captivate and motivate the laboratory staff so quickly? First and foremost, it was his innate ability to see the big picture. It was as if he could see the entire world in all directions at the same time. He was able to assess the full complexity of developing computer technology, to clearly formulate basic ideas for its design, and to outline present and future goals in the field. His personal unshakable conviction in our scientists and their ability to accomplish anything they set their minds to, was infectious.

Personal Reminiscences of Viktor Glushkov, cont'd. January 3^rd 1982

The computing machine designs of that time were based on engineering intuition, so I had to learn the principles of computer design on my own. As a result, I became intimately familiar with the inner workings of a computer. Since then, the theory of computing machines has become one of my specialties leading me to transform the art of computer design into science. Naturally, the Americans were working on the same issues, but they achieved similar results later. The theory of automatic machines as the basis for computer design was not sufficiently developed back then. Apparently, the first people to express the idea of the possible applications mathematical logic in the design of technical devices were Claude Shannon in the United States, and V.I. Shestakov and Mikhail Alexandrovich Gavrilov in the USSR. They applied the simplest formal mathematical logic to a design of telephone switching circuits. It appeared to also be useful for simple electronic circuits. Therefore, in the post-war years when digital computing technology started developing, attempts were made to apply formal mathematical logic for solving tasks in computer circuitry. I began to work on this problem and organized a seminar on the theory of automatic machines. The essence of one my first projects was the discovery of a much more elegant, algebraically simple and logically clear concept for Kleene's automatic machine and I obtained all of Kleene's results. But more importantly, even beyond Kleene's results, I was developing a theory aimed at real tasks of machine design. At the seminar, we discussed issues in Kiev's design and were able to see which parts of my theory would work and which parts would not.

Alexander A. Letichevsky, another of Glushkov's contemporaries, recalled:

Julia Kapitonova became the "soul of the seminar." She was Glushkov's favorite student. Viktor Bondarchuk and I were also faithful participants in the seminar. It was a romantic period: a new science was being born right before our eyes and we were expanding its frontiers with our own hands. We were proud when we managed to solve the tasks our teacher gave us. The theory of automatic machines was not chosen accidentally by Glushkov. It was a well-planned technical move. As an algebraist, Glushkov saw that the concept of automatic machines, coming from Stephen Kleene and other authors in the well-known Automata Studies [published in 1956, Princeton, N.J. under the editorship of Shannon and McCarty; translated into Russian the same year under the editorship of Alexei A. Lyapunov] presented a powerful mathematical model for a discrete information processor. This model could be studied by application of a rich array of contemporary mathematics. At the same time, the expansion of applied theory on the basis of rational mathematics could attract the attention of engineers who lacked sufficient knowledge of mathematical theory to design devices containing memory elements. In addition, the theory of automated machines could become the basis for the design of cybernetic system models with various applications.

Personal Reminiscences of Viktor Glushkov, cont'd. January 3^rd 1982

I was directing a large team for the first time, so I had to define some organizational principles. I didn't write them down, but I followed them consistently, and it always led to success. In particular, I used the principle of unity between theory and practice with a new twist: when undertaking a large design project, consider both the present and future goals. In a new science like cybernetics, one must always break down a long-term project into smaller, more manageable pieces. Each stage would then become a step toward the ultimate goal with its own timeline. At the same time, its completion would yield an independent result with distinct benefits.

I have a habit that developed in my early childhood: when I am interested in a new field of study, I like to do preliminary research before beginning to work in it.. Prior to my arrival in Kiev, I was involved with topological groups. I had a clear idea of what can be expected from any scientist dealing with this problem. That is, I easily felt the rhythm in problem-solving, and I knew I had to stay a step ahead. The feeling of excellence was necessary for me to consider myself a specialist. For three years, I'd been racking my brain trying to solve the basic theorem on Hilbert's generalized fifth problem. My subconscious worked even when I was sleeping. Sometimes at night, it seemed like I had solved it, but then in the morning I would get up, sit down at my desk and alas, somewhere, somehow there was a logical mistake. A continuous three-year onslaught ended in 1955. My wife and I went on a hiking tour of the Caucasus. While climbing an ice field on Mt. Kazbek, I was struck by an idea that enabled me to solve Hilbert's problem. However, I had gotten so used to the fact that there was always an error in my reasoning that at first didn't believe I had finally solved it. I looked for errors, but couldn't find any. On the way back, riding on the train, I quickly wrote down the solution and spent the next six months polishing it. It came to sixty pages and proved only one of the theorems. Up until that time, no one in the world was able to come up with a shorter proof. This work brought me recognition among mathematicians and a tremendous creative satisfaction, so to speak.

I believe that I was able to coordinate the work at the laboratory not because of good organizational skills, but because I can see things in the larger context, and I can direct research, set goals, motivate people and guide them through the process. That is my saving grace. Somehow, I learned something; I even formulated several organizational principles along the way, but it is not my strong suit. I like to spend my free time proving theorems. That's what I truly love and feel most comfortable doing. Organizational work, on the other hand, is a burden to me. Sometimes it becomes interesting, and I get absorbed in it, but only because I enjoy seeing things through to the end.

Personal Reminiscences of Viktor Glushkov. January 4^th 1982

In December 1957, the government and the Presidium of the Academy of Sciences of Ukraine made an official decision to establish the Academy's Computing Center. By that time, our staff consisted of more than 100 people. The Ukrainian Academy of Sciences allotted financial support for the construction of the Center on Lysogorskaya Street, and a block of apartments for the staff. In the beginning, the Computer Center was to be supplied with three computers: the Kiev, the MESM, and an Ural-1. The latter machine was the latest one manufactured. The building was designed to house 400 workstations and had three large halls in which to install the computers. In 1959 our staff moved from Feofania to Kiev, but the building was not finished in time. That was an interesting period in the history of the Computer Center. According to the technical requirements, the premises where the computer equipment was to be installed had to be clean and air conditioned. Yet, we had to do the final checks and launch the Kiev computer before the building had a roof. What really helped in this difficult process was the enthusiasm of our young staff. Of course, the building was finished shortly thereafter.

The Kiev computer played a significant role in our work, though it was never put into serial production. The second Kiev computer was bought by the International Institute of Atomic Research in Dubna, near Moscow. During 1956–1957, nuclear research was booming. Therefore, collaboration with this Institute helped and taught us a great deal. On one hand we were on the forefront of science; on the other, we were learning how to cooperate with industrial enterprises. During this time my main task was to establish the fundamentals of computer theory, which I completed in 1961. The research was extremely intense and time consuming. I spent whole days at the Institute; during the nights, I poured over books and articles, and often wrote until five in the morning. Working so hard and for such long periods of time without rest, took a toll on my health. At the beginning of 1963, I suffered a brain seizure and was briefly hospitalized. After that, I was forced to change my working habits and take care of my health.

Glushkov's wife had tried to step in and make her husband pay more attention to his health. But she did not manage to convince him:

He worked eighteen to twenty hours a day. He would get so busy at work that he forgot to eat. When he came home, he rushed to his desk and worked until the wee hours of the morning, sometimes until dawn. Viktor did not listen to warnings about the dangers of overexerting himself, but his reasons were understandable. In a short period of time, he was supposed to learn all of the latest trends in his scientific field. In addition, he was responsible for a large staff of scientists this time, not just for himself. A lot of organizational difficulties came up and everything new required a great deal of effort to break through. After his treatment was over and he returned to work, he changed his schedule a little bit, but he still did not rest enough. He had this note at home; it was always under the piece of glass covering his desk: "Today is the first day of the rest of your life. Don't waste it."

Personal Reminiscences of Viktor Glushkov cont'd. January 4^th 1982

My book, The Design of Digital Automatic Machines, was published in 1961 and became the basis of a new scientific trend in our Institute and probably played a positive role in the nation as well. In 1964, my book received the honorable Lenin Prize. I wrote several other books during that time. The monograph Introduction to Cybernetics was finished while I was still in the hospital; it was published in 1964. Both of those books were also published in the United States and in many other countries. At the same time, I also wrote a theoretical article, which established the basis for many works on the theory of automaton with applications for algebraic theory of automatic machines. This article, ‘Abstract Theory of Automatic Machines,' was published in the magazine Successes of the Mathematical Sciences, and therefore available to a wide circle of mathematicians. It was republished as a brochure in the German Democratic Republic and other countries. Influenced by this article, many of our algebraists became involved in the theory of automaton. I have to point out the special feature of our school: we strived to be more practical rather than theoretical.

Simultaneously with theoretical research, we pursued the creation and application of computer technology in Ukraine. At that time, enterprises used simple analog computing devices for computer-aided manufacturing control. For each technological process, they created its own separate device, especially for the ones that could be described by simple differential equations. In 1958, at the Computing Conference in Kiev, my proposal to build a multi-purpose digital control computer with a transistor was met with hostility. Moscow scientists, led by the academic Vadim Alexendrovich Trapeznikov and many other Soviet computer science specialists, had the same negative attitude towards it: they still believed that computers should be based on electronic vacuum tubes. These devices required huge amounts of space and proper air-conditioning, which conflicted with the realities of manufacturing and on-site control of technological processes.

At that time, Boris Malinovsky was one of the first to do research work on semiconductors, which was very helpful to us in developing a Universal Control Computer. We outlined the basic ideas, which later became the governing philosophy: the computer had to be built with semiconductors, needed to be portable, have reliable protection, and a word length of 26 bits – long enough for technological control of the majority of the processes. But the main idea was a special interface unit (a set of computer controlled analog-to-digital and digital-to-analog converters), that could connect the computer to the technological processes.

Malinovsky was the chief designer of this machine and I acted as the scientific advisor. From the time this idea was presented at the June 1958 conference, it took only three years to begin serial production of the computer (in July of 1961) and to install it at some facilities later that year. As far as I know, this is still the shortest time frame for the design and installation of new technology.

The Universal Control Computer, later named the Dnepr, was a collaborative effort with several Ukrainian enterprises to incorporate the control of complex technological processes. Jointly with the Dzerzhinsky Metallurgy Plant – in the city of Dneprodzerzhinsk – we analyzed issues in controlling the process of steel-casting in Bessemer converters. We solved similar problems of column carbonization with the workers at a soda factory in the city of Slavyansk. I also initiated an experiment – the first of its kind in Europe – to control those processes remotely for several days in a row. In addition, we conducted research on Dnepr's applications to computer-aided welding processes at a plant in the city of Nikolayev. Malinovsky, Vladimir Illich Skurikhin, Gleb Alexandrovich Spinu, and others were deeply involved in these projects.

We learned later on that American scientists started developing a universal semiconductor control computer – analogous to Dnepr – earlier than us, but began producing it at the same time, in June 1961. This was a point when we managed to catch up with the Americans in one very important scientific area. It should also be noted that our computer was the first Soviet semiconductor device [with the exception of computers for military purposes]. Dnepr turned out to be a very sturdy machine, able to withstand extreme weather conditions, vibration tests, etc. It also set a longevity record – Dnepr was manufactured for ten years, from 1961 to 1971, while typical computer equipment required serious modernization after five to six years in production. After much debate during the joint Soviet-American Soyuz-Apollo space flight, this computer was chosen to control the display screen at the Flight Control Center. Dnepr was also exported and employed in many Socialist countries.

It should be noted, that the seven-year-plan (1958–1965) for Soviet industrial development in Ukraine did not include the manufacturing of computers. The first Dnepr computers were made at the Radiopribor factory in Kiev. At the time of Dnepr's design, we initiated the construction of a Control Computer production plant (in Russian: Zavod Vuichislitel'nikh i Upravlaushikh Mashin, or VUM, now called the Electronmash) that was supported by the government. Thus, the heroic period of our development came to an end. I call this time heroic because we worked in difficult conditions and were always expected to perform above and beyond the scope of our jobs.

I frequently wrote and spoke about this. Unfortunately, my organizational efficiency coefficient (as I once calculated) did not exceed four percent. What does that mean? It means that in order for a problem to even be considered by the government, I had to speak with twenty-five officials. However, after the success of Dnepr, I was generally well received and with much less skepticism than before.

Research on control computers did not stop with Dnepr. Let's skip forward to note the following basic design projects: in 1967, in cooperation with the Institute of Cybernetics, the VUM plant started manufacturing a new model, Dnepr-2. This computer contained complex multilevel interrupt system and an effective real time operational system. Unfortunately, its productions soon stopped. In 1976, a terminal processor control computer, BARS [in Russian: Bazovaya Aparatura Razrabotchika Sistem, basic system creation apparatus. The noun ‘Bars' also means 'Snow Leopard' in Russian], was designed by Vladimir Skurikhin, Anatoly Morozov, and others. Its design received the Golden Prize at the International Exhibition in Dresden, and it was used at several industrial plants. In 1977, the M-180 computer control complex was created and put into production, along with a system of technical interfaces designed by Malinovsky, Pavel Sivachenko, Alexander Palagin, Yuri Yakovlev, and Vladimir Reutov.

Personal Reminiscences of Viktor Glushkov. January 5^th 1982

In 1962, our Computer Center was reorganized and renamed the Ukrainian Academy of Sciences Institute of Cybernetics. The Institute grew quickly and we were involved in many areas: probabilistic automatons, functional reliability of automatons, economic and energy saving, and interference resistant coding. The research started shifting from finite-state to infinite-state automatons. We discovered the connection between the theory of automatons and the theory of formal grammar. We implemented new methods of analysis and automaton design; besides me, Letichevsky and Kapitonova were actively involved in this research. Their work gained wide recognition.

In 1959, I began designing a computer that would perform engineering calculations. This project started with the design of a digital calculating automaton. In 1963, we launched serial manufacture of the Promin computer. By that time we understood that we needed a design bureau, which was established in 1963, but its prototype appeared much earlier at the Institute of Cybernetics. The staff that created the Promin later joined the design bureau.

Promin's manufacture began at the Severodonetsk computer plant because the Kiev plant was still under construction. Technically, the Promin had a number of innovations, in particular, memory on cards coated with metal. But most importantly, it was the first widely used computer with step-by-step micro program controls. Unfortunately, we were not able to obtain an international patent for the new scheme of control. Back then, the Soviet Union was not a member of the International Patent Union and we could not obtain the copyrights. Later, the step-by-step program control was used in the computer Mir-1 for engineering calculations. In 1967, at an international computer show in London, the Mir-1 computer was purchased by the IBM – the biggest computer company in the United States and the exporter of eighty percent of the computer equipment in the western world. It was the first – and unfortunately the last – purchase of a Soviet electronic computer by an American company. As it became apparent later, IBM did not buy the computer for calculation purposes, but to prove to their American competitors who had patented the principle of step-by-step microprogramming in 1963, that the Soviets had known about this principle long ago and had used it in serial production of their own computers. In fact, we put it to use even earlier, in the Promin.

A new and upgraded computer, Mir-2, went into production in 1969, followed by Mir-3. These computers were unmatched in their analysis conversion speed. For example, Mir-2 successfully competed with larger, standard structure all-purpose computers, which exceeded Mir-2 in nominal rate of speed and had a hundred times the memory capacity. The Mir was the first Soviet computer to implement a dialog mode, using a display with a light pen. Each of these machines also represented a large step toward designing computers with artificial intelligence. It was a strategic breakthrough in the development of computers.

What was the difference between Mir and other computers? We considerably upgraded the machine language. However, back then the popular point of view was that machine language must be as simple as possible and the rest would be done by software. We were even mocked for our efforts to develop different computers. The majority of computer scientists in the world believed that it was necessary to develop computer-aided programming, that is, to create software that would help produce other programs. Our colleagues Korolyuk, Yushenko, and other scientists were engaged in this field and were the first in our nation to suggest an effective address language for the Kiev computer, creating the so called ‘programming programs' (translators) for other computers. But I did not take part in the work.

In designing the Mir machines, we had tackled a daring problem – to match the machine language as close as possible to the human language, and here I mean mathematical nonverbal language, though later we made attempts with normal human language. So, we created ‘Analytic,' a special mathematical language, supported by an internal interpretation system.

Mir computers were used in all regions of the Soviet Union. Their creation became an intermediate stage in research aimed at the development of artificial intelligence, since the intelligence realized in them was still fairly primitive. It also looked very impressive when a machine quickly solved independent and dependent integrals, while not many professors of mathematics were able to solve them. In addition, the machine found substitutions, not just the easy ones from tables, but the difficult ones as well.

In 1966, Glushkov and Rabinovich published the first article in the world on improving computer efficiency by simplifying the human-machine interactions: "On a Few Problems of Algorithmic Structures in Computer Systems," (Cybernetics in Service to Communism, Moscow: 1966).At that time, those were "revolutionary views" acknowledging that the direction of computer development was shifting. The first battle for this new ideology had already occurred (in 1962) at an international conference on computer development in Kiev. Participants came from Bulgaria, Hungary, Poland, and Czechoslovakia. Glushkov was scheduled to talk, but suddenly fell ill. Despite his high fever, he decided to make his briefing because he believed that the conference was very important. Unfortunately, his illness prevented him from speaking with his characteristic zeal, which would have electrified the audience. After the presentation, he was bombarded with hundreds of challenging questions. The renowned Moscow specialist, Mikhail Romanovich Shura-Bura, sarcastically remarked that if one were to realize all of Glushkov's suggestions, the proposed computer would be larger than the building in which the conference was being held. At the end, things settled down, but his opponents clung to their opinions.

The importance of integrating artificial intelligence into computers was recognized in 1963 at a rather modest symposium, organized by the Cybernetics Institute and Uzhgorod University in which Lebedev, Glushkov, Mikhail Sulim – chief administrator of computer technology in the Ministry of Radio Production – and several others took part. In general, the atmosphere in which we discussed our proposals for computer architecture was friendly, and our critics remained quite benevolent. Another camp of mathematicians was present, and I remember that even though our discussion was emotional, everyone remained quite businesslike. Lebedev liked our proposals and noted that several of them overlapped with the ones used in BESM-6's development. As the result of the conference in Uzhgorod, our proposals were discussed and approved, and all of the participants expressed their recommendations regarding the direction of computer development. The "high-level side" – Lebedev and Glushkov – finally agreed that the Institute of Precision Mechanics and Computing Technology would work on creating supercomputers, while the Ukraine Institute of Cybernetics would take on the development of smaller and more specialized computers.

Returning to Kiev, Glushkov began designing Mir-1 with renewed energy. Within two weeks he had drawn up the preliminary plans, identifying the main architectural and structural contours of the machine. It contained a series of original solutions, which would serve as the basis for an invention patent. A close working relationship between scientific colleagues at the Institute of Cybernetics and the scientists and engineers of SKB-245 led to brilliant results – the Mir computer family was quickly developed and put into serial production, receiving high marks from its users. Its creation was a giant step in the development of artificial intelligence in small computers.

Unfortunately, the potential of the Mir computer line was never fully realized. During my 1979 presentation in Novosibirsk on the integration of artificial intelligence into computers, I heard the academician Andrei Ershov criticize the Institute of Cybernetics by saying: "If you had not stopped upgrading the Mir family, the USSR would have had the best personal computers in the world."

Personal Reminiscences of Viktor Glushkov, cont'd. -January 5^th1982

Computer architecture is following a special path because new ideas still come from people. In our work on computer architecture, I took consecutive departures from the well-known Von Neumann principles. For example, the sequential structure of language: the fulfillment of one command after another; a command-address principle; a command containing address operands; commands which are saved as operands in memory; the simplest system of commands; and the simplicity of machine language. There are other principles, but these are the main ones. The appearance of such principles was not surprising. In the era of vacuum tube computers, when each arithmetic bit in the structure required a minimum of one triode, a simple machine with simple commands was necessary.

Even back then, I anticipated the development of microelectronics to the point where all hardware components would be made at the same plant in a streamlined process and become very cheap. To prepare for this, I proposed for our physicists to construct a physical medium for the creation of a computer. In this situation, Von Neumann's principles were not applicable. I suggested a complex machine language as one of the new principles, because compilation systems were becoming more intricate. It was necessary to simplify the programming process from both ends – for languages and for compilers – in order for the machine language to simulate the input language. After partially integrating this idea into the Mir computer series, we continued to develop it in accordance with the principles of progressively complex machine languages, to get closer to human language. My goal was to be able to speak directly with the computer and issue commands in our language.

In order to have a conversation with a computer in a spoken language, the logical reasoning components must be automated first. That is the easiest step since some of the formalisms needed are already known. But further analysis showed that the classical mathematical logic does not account for all of the necessary steps. Therefore, the task of constructing a practical mathematical logic was put forward and was successfully resolved. It was a pivotal point to realize that a mathematical proof can be designed like a program, based on language. When we are able to accomplish this, we can then introduce such a language into the architecture of the machine. Computer-aided proof of mathematical theorems is my ultimate dream. It is the basis of my ideas for new computer architecture; for the kind of computers that are capable of complex creative processes and deductive reasoning. In other words, computers, that build other computers. That is where the new ideas for computer design will come from. However, only the people who work with both computers and artificial intelligence will be able to build them. That is our strength.

At the end of the 1960s, Glushkov began supervising the development of the Ukraine computer and appointed Rabinovich as the chief designer. His chief assistants were Alexander Stogny and Ivan N. Molchanov. It was the next step in the departure from von Neumann's principles towards integrating artificial intelligence into computers. This time it was tied in with the development of high-performance, universal computers and a schematic realization of high-level language.

The development of the Ukraine computer was an important landmark in the growth of Glushkov's scientific school. The ideas embodied in the Ukraine machine surpassed many of the ones used in the American computers in the 1970s. Besides making the machine language more complex, we tried to switch from von Neumann's principles of sequential command execution to a multi-command mode. We encountered many obstacles until someone came up with the idea of a macro-conveyer and we were finally able to make a multi-command computer with many command streams and data channels.

Glushkov proposed a macro-conveyer principle based on the idea that each processor was given a separate task during every step of the computing process, which allowed it to work independently for a long time without the interference from other processors.

In 1959, at the Soviet All-Union Conference on Computer Technology in Kiev, Glushkov spoke about the idea of a brain-like computer structure that could be realized when the designers were able to integrate not thousands, but billions of elements with practically limitless connections between them, into a single system. There would also be a confluence of memory and data processing, a system in which data would be processed throughout the memory with a highest possible degree of parallelism in all operations.

At the 1974 International Federation of Information Processing (IFIP) Congress, Glushkov presented a paper on the recursive computer, based on new principles of computer system organization. He argued that only the development of new non-Von Neumann computer architecture, based on a current level of computer technology, would solve the problem of creating a supercomputer with unlimited growth in productivity and progressively more sophisticated hardware. Unfortunately, further research showed that a comprehensive realization of the construction principles of recursive computers and brain-like structures was beyond the level of electronic technology at that time.

"It is imperative to find a reasonable solution in order to transition from the Von Newman principles of computer design to the brain-like computer structures of the future," Glushkov said in his report at the Novosibirsk Conference in 1979. Glushkov promoted this idea as the basis of the original structure for a high-performance macro-conveyer computer, and worked on this even during his tenure as Vice-President of the Presidium of the Ukraine Academy of Sciences.

In 1981, a well-known nuclear weapons designer, academician Yuli Borisevich Khariton, visited the Institute of Cybernetics at the Ukrainian Academy of Science. He had become very interested in the unusual macro-pipeline computer because of its computing speed, which was significantly faster than any other machine and it greatly reduced valuable processing time for many important projects.^[2]

Glushkov understood the importance of Khariton's visit for the future of macro-pipelined computers and the institute as a whole. By this time, Glushkov was already terminally ill with a rare brain cancer – medulla astrocytoma – that made talking very difficult, plus his speech was constantly interrupted by a nasty cough. Nevertheless, he received Khariton himself. Glushkov was brimming over with enthusiasm for the idea of a powerful Soviet supercomputer. He truly believed that its completion would greatly help our physicists. Glushkov did not live to see the realization of his ideas in the Unified System (ES)-2701 and ES-1766 macro-pipeline scientific computers.

According to a Soviet government commission that evaluated the project, these computers were unique in the world. With its full complex of 256 processors, the ES-1766 was estimated to perform at a half billion operations per second. The ES-2701 and ES-1766 design plans were transmitted to the Calculating-Electronic Machines Factory in Penza, Russia, for serial production in 1984 and 1987, respectively. These machines rivaled the best American computers, were extremely powerful and sought after for scientific applications. Unfortunately, very few of them were ever produced.

Up until this time, the Soviet Union put many computers into serial production via the Institute of Cybernetics at the Ukrainian Academy of Science and the SKB. These included series of mini-computers, specialized computers, and keyed-program computers: SOU-1, Neva, Iskra-125, Mria, Chaika, Moscow, Scorpion, Romb, Orion, Express, Pirs, and others. The Institute of Cybernetics, in conjunction with the S. P. Korolev Scientific-Manufacturing Company, created a whole complex of microprocessor computers: the Neuron series and the debugging systems SO-01 and SO-04, developed by Malinovsky, Palagin, and Valery Iosifovich Sigalov. They also took part in the design of the first Soviet microcomputer Elektronika-S5, manufactured at the Svetlana Electronics Plant in Leningrad.

Modern computers were impossible to build without a system of computer-aided design. The Institute's extensive research enabled it to create a series of unique systems: the Proekt family -- Proekt-1, Proekt-ES, Proekt-MIM, and Proekt-MVK -- for computer-aided hardware and software design. The Kiev computers were initially employed for this process, and later the M-20, M-220, and BESM-6. The Proekt-1 was a specialized, programmable device with its own operating system. Glushkov, Letichevsky, and Kapitonova pioneered the optimized automation of algorithm design for it. The Proekt series was experimental: they laid down the groundwork for software and hardware design, and were subsequently utilized by dozens of organizations and computer scientists.

The Proekt-1 system was used in the automated project design of BIS (in Russian: Bolshiye Integralniye Skhemii, or Large Integrated Circuits), with the help of special electro-ionic technology. In Vitaly Pavlovich Derkatch's department at the Cybernetics Institute, the Kiev-67 and Kiev-70 computer installations employed these technologies. The Proekt's computer-aided design system had a communication interface with Kiev-67 and Kiev-70 that controlled an electron beam during the chip base processing in real time. In 1977, Glushkov, Derkatch and Kapitonova received the Soviet Union State Prize for their work on this project.

Computer-aided programming was one of Glushkov's main interests, and he viewed the development of algebra for algorithmic languages as the path to perfecting this technology. In studying this problem, he considered not only the general mathematical principles, but also philosophical concepts. Comparing numerical and analytical methods of task solution in applied mathematics, Glushkov asserted that the development of general algorithmic languages and structures for such languages would allow them to be widely used, just like the analytical expressions have become in today's computer programs. Today, the differences between analytic and general algorithmic methods are disappearing, and computer models are becoming the basic platform for the development of new mathematics.

Personal Reminiscences of Viktor Glushkov. January 6^th 1982

Simulation of sight and hearing are important research components in the field of artificial intelligence. The most important element, of course, is sight; it provides the greatest amount of information.

From the very beginning, I wanted to automate robotic motor functions. I began with the task of creating an automatic arm on a handcart, which could move along a control panel and change the position of the tumblers, pull-switches, turn knobs and so forth. It would be equipped with primitive sight, capable of perceiving only the position of device indicators or units on a scale. Unfortunately, I could not find a person who loved working with his hands. I was working on this task in 1959, when no one had even mentioning robots. If we had had good workshops and mechanics, then by 1963 we could have been the first in the world to develop a mechanical arm. Regrettably, we did not succeed.

At the same time, we began working on phrase recognition in Russian, or what is now called semantic networks. Alexander Stogny and Letichevsky were both involved in this project and we achieved good results. I developed the algorithms, and Stogny wrote the programs. The algorithm built a semantic network based on the sentence flow, identifying which of the words corresponded with which. For example, although the sentence "The chair stands on the ceiling," is grammatically correct, semantically it is not.

When Stogny changed the direction of his work, I also had to stop. We simply had to let the project go. Even though we needed to link the algorithm with an advanced computer, there weren't enough people to complete the work, and I could not spend all of my time and energy on semantic algorithms. Nevertheless, when I wrote a paper on this subject for the 1961 IFIP Congress in Munich, it became a sensation – the Americans had nothing like it at the time. After that I was selected to be on the committee for the International Federation of Information Processing.

G.L. Gimmelfarb, a retired employee of the Cybernetics Institute, recalled:

The Kiev computer became the first machine in Europe with a system of digital imaging that was capable of modeling intelligent processes. Two innovative peripheral units were attached to the Kiev and enabled it to simulate the simplest algorithms for learning image recognition and single-purpose the prototypes of today's monitors – were found only in the United States.

The Kiev computer, under the direction of Glushkov in the late 1950s and early 1960s, fulfilled several research tasks for artificial intelligence: studies on recognition of simple geometric figures, on prototypes of automated readers for handwriting and text recognition, on tracking movement, on simulating the behavior of a group of automatic devices in the process of evolution, on the automatic synthesis of functional computer schemata, and others. That is how Glushkov got involved in both the theory and practice of simulating intelligence during the infancy of computer technology, when many people perceived computers simply as "big counting machines."

Later Kiev's input-output image units were modernized and carried over to the BESM-6. With these units many types of work were carried out, including the digital analysis of photographs of real objects, particularly for the discovery and compression of trace physical particles in bubble chambers; also the tracking, recognition and compression of movement of different means of transport, the recognition of text messages, and others.

The experience gained in creation and use of input-output image units gave rise to the development of the first Soviet simulator in the 1970s for the modeling of intelligent hand-eye type robots. At the core of the simulator was the BESM-6, a television system for image input and an electro-mechanical manipulator with six degrees of mobility, connected to the computer through an M-6000 mini-control computer. Glushkov was greatly interested in this work because he considered robotics to be one of the most important practical directions for using the methods and means of artificial intelligence.

^[2] Khariton was the scientific director of Arzamas-16 (now Sarov), the Soviet secret nuclear weapons design laboratory, for over 40 years.