Soviet calculator “Vilnius”; image 7md.lt
Soviets managed to go head-to-head with the US in terms of nuclear weapon development. They also managed to stay ahead of the US in terms of rocketry and to hold their own in the space-race. However, on the technological frontier of computing they have fallen horrendously behind. Why was that so?
This is a complex issue and there are multiple causes for this. Computers are essentially information technology and as such were tightly controlled in the Soviet Union. There were almost no hobbyist magazines or other spaces were people with interest in computers could exchange know-how, ideas or showcase their work. Furthermore, a perpetual scarcity of essential parts (transistors, chips, capacitors, etc.) existed in the Soviet Union, which made taking up such a hobby very difficult. Consequentially, a strong hobbyist community which started such great enterprises as Apple, HP or Microsoft in the US never developed in the Soviet block. It was without a doubt detrimental to the Soviet computer industry, but in this post, I’d like to explore another side of the story, how personal ambitions, and inter-institutional rivalries effectively sunk the Soviet computer industry.
Early days – 1930s
It was WWII that kick-started the development of first electronic computers. Before that computers were analog – instead of discrete electrical impulses they relied on continuous change of a certain medium. In the US, the most advanced such machine was a Differential Analyzer – a device which had a number of gears and levers and outputted a continuous graph. In their most advanced configuration the differential analyzers were able to sort any differential equation.
In those days, the Soviets were not far behind. In 1936 a railway engineer Vladimir Lukyanov developed Water Integrator. It was an analog computer, which used water passing through different sized containers to simulate thermodynamic exchanges of energy in complex systems. These exchanges were fairly complex and could be described by 3rd order differential equations. An operator would prep the system, let the water flow through and record the water level at the output container. After several runs with different parameters, different data points could be plotted producing essentially the same type of output as as the Differential Analyzer.
Though developed for a very niche problem – to solve the problem of cracking concrete used for the railway projects, the water integrator was widely adopted in other industries with over 150 Water Integrators being built. They were used in the Soviet Union for large scale modelling until 1980s (<- That is not a typo; really 1980s).
Cybernetics
So what happened? It would be rather impossible to understand without setting the scene a little. Until 1950s computers and computers were a very niche pursuit with very niche applications. Good for automating large amounts of repetitive calculations, but not much more. However, after the development of the first electric computers that started to change somewhat. In the US the development of computers were led by a group of mathematicians who had a very broad outlook and were able to foresee wider applications of these machines.
For instance, Norbert Wiener, a mathematician at MIT, held a series of talks with physicists and medical students and later collected the notes from these discussions in to “Cybernetics: Or Control and Communication in the Animal and the Machine”. The book basically outlined the principles how self-regulating and self-collecting systems function. The book stipulates that the same essential principles are applicable to the functioning of both computers and human brain. And that eventually, it will be possible to construct thinking machines.
Just a few years later, one of Wiener collaborators, Claude Shannon, constructs Theseus a maze-solving mouse robot. Essentially, a collection of connected electromagnetic phone relays this robot was seemingly capable of intelligent behavior, learning and memory, which suggested that general AI is just around the corner.
This was the beginning of the first AI hype and the zeitgeist was perfectly captured in an article “Can we build a superhuman?” in the 1950 Time magazine, with one of the most advanced computers at the time – Harvard Mark III on the cover. The article drew several parallels between computers and human brain and asked whether computer operations could be considered “thinking”.
Литературная газета article
At the same time (1950) NATO was formed, which led the Soviets to intensify their anti-american campaign and many journalists started actively looking for any excuse to go America-bashing. One of these was Boris Agapov, scientific news editor at arts-and-culture magazine Литературная газета ( Literaturnaya Gazeta ). He published an article heavily criticizing cybernetics, a term borrowed from Wiener’s book. Article mocked stupid Americans for wishing to replace class-conscious workers with machines and robots. It is important to note that the entire critique was based on the Times article and not the actual Cybernetics book, which Agapov never read.
For some reason, Agapov’s article in Литературная газета became widely influential in at least two ways. First, it made the term Cybernetics stick. It eventually became a wide umbrella term for anything related to computer-hardware, computer-science, networking or information theory. Second, it set the tone for the negative reception and denouement of cybernetic in the Soviet Union.
At this critical time for the development of computers Soviet denouncement of Cybernetics meant that research projects were stopped, funding was not allocated and scientists generally feared to propose anything that could be considered cybernetic related in fear of losing their jobs or worse.
It might be hard to understand how could an article in a arts-and-culture magazine have such a profound effect on the development of computer science, but such occurrences were quite common at the time. This occurred due to to features of the Soviet system: high degree of centralization and absence of checks-and-balances. High degree of centralization meant that a very small group of people made all the high-impact decisions. The absence of checks and balances meant that the system was very vulnerable to bandwagoning and ‘mood-swings’. You essentially needed to convince a few key people to cause a radical change in policy.
Cybernetics in service of Communism
Cybernetics was shunned until the death of Stallin in 1953. Then slowly things began to change. This change was mostly brought by the increased demand for computing in the space program as well as in various weapons programs. Progress was made and under the supervision of Ministry of Radio Technology (Министерство радиопромышленности) started production of domestic computers in two factories in Minsk (produced Minsk computer series) and in Penza (produced Ural series).
Soviet economic system relied on central planning, which meant that all the information pertinent to sector of the economy was collected in one place, processed and decisions made on the collected data. There was a lot of data that needed to be analysed and a lot of decisions to be made. All of that could be tremendously aided by a computer. So as post-Stalin Thaw progressed, Soviet mathematicians and engineers began to push their case in earnest and managed to convince upper Party eschalons of the need for rapid computerization. This culminated in the 22nd Congress of the Communist Party, which proclaimed “Cybernetics in service of Communism” – a full embrace of computerization.
However, computer industry was in a very big hole due to absolute lack of progress in the critically important early years. The problem was further amplified by the closed nature of the Soviet system – it was simply not possible for Soviet engineers to work in the West for a few years and then apply their skills upon returning home. Instead, everything had to be learned by trial and error, which made catching up difficult.
A good illustration of this is the computing in the Soviet space program. Manufacturers of Vostok and Soyuz spacecraft submitted specifications (desired capabilities as well as desired size and weight specs) to all computer manufacturers and none could actually produce an on-board computer that could perform needed calculations and still fit inside in the spacecraft. As a result Yuri Gagarin blasted into space in 1961 without a computer on board. The same applied to early Soyuz capsules throughout the 1960s. This meant that the spacecraft re-entry and other calculations were performed on the ground and then relayed to the spacecraft. It was much more dangerous, as any radio communication disruption could have been fatal.
Failing
The time is 1960s. The Soviets are quite far behind the West in computers and computerization. The demand for computers is great keeping in line with the “Cybernetics in Service of Communism” directive. The resources are scarce and current computer production factories cannot meet the demand. In a market-economy that would trigger increase in the price of the critical resources and would channel them to the sectors with greatest demand. Also, more producers would pop-up to meet the increased demand. In the planned-economy such supply-demand driven self-regulation was not possible and more producers could not enter the market without the approval from the center.
Such decision was made in 1965 when a new Ministry of Electronics Industry (Министерство электронной промышленности) joined-in the computer production. However, this was too little tool late. Targets for computerization began to slip and everything was falling behind schedule. In this environment in 1966 the Soviet leadership made the only “sensible” decision – to stop wasting time and resources on the domestic R&D and to concentrate on simply cloning Western machines.
However, even cloning Western computers was not easy. Part of the problem was that the Soviets refused to follow IEC and IEEE standards for such aspects as pin spacing on micro-chips and instead insisted on their own standard. This meant that Soviet hardware was incompatible with international standards. In other words all components even the most trivial ones had to be produced domestically and no components could be exported. In short, this prolonged the delays and made parts more expensive. Consider the following, IBM launched System 360 in 1964 – it was the first modular mainframe computing system, meaning that new components could be added to the existing hardware without needing to replace the whole thing. Soviets decided to clone the hardware and could do so, because IBM had patented everything in great detail. However, due to aforementioned problems it still took 5 long years to produce a partial clone.
Furthermore, due to generally poor quality of Soviet goods these computers were continuously plagued with problems. I remember numerous stories how the payroll calculations produced by MINSK computer had to be checked by hand, because poor quality paper used punch cards meant that they could not be read correctly and introduced errors into the calculations. Because of this, when it was essential that the system runs flawlessly Soviets often simply used American machines. For instance, an IBM computer was bought to handle the ticketing and hotel reservations for the Moscow Olympics in 1980.
OGAS System
Among the general failings of the Soviet computerization, it is particularly worth singling out the OGAS project. OGAS ( ОГАС, Общегосударственная автоматизированная система учёта и обработки информации) or a “National Automated System for Computation and Information Processing” was an attempt to create a Soviet Union wide network of computers which would crunch the economic data and help to facilitate the central planing process.
The idea was to establish the main computing facility on Moscow and connect it via a network to ~200 regional computing centers. Additionally, up to 20 000 terminals would be established in various places, enabling wide access to the network. Such idea was proposed in 1962 by Viktor Glushkov (Виктор Глушко́в ), director of the Institute of Cybernetics in the Ukrainian Academy of Sciences. Given that the granddaddy of the modern internet, ARPANET, did not become operational until 1969, his idea was really cutting edge. Some plans even theorized that the OGAS system could be used to log any exchange transaction that occurred in USSR and thus enable money-less economy – a major achievement en route to full communism.
The project kicked-off in 1962. Well respected and trusted Viktor Glushkov stood at the helm and the National Academy of Sciences were firmly behind the idea. However, the plan was dead in the water by 1970 with little to show for. What happened?
A few things. First, nobody really wanted such computer network. Soviet system was unbelievably corrupt. Many (factory managers, low level bureaucrats, national civil servants, etc.) felt threatened that they would not be able to “cook the books” and that such monitoring system would expose just how dysfunctional the Soviet economic system really is as well as the extent of graft and other forms of corruption. Second, the goal of creating money-less economy directly threatened several key stakeholders, mainly the Ministry of Finance, which would have become obsolete if the project was successful. It was rumored that that the head of the finance ministry, Vasily Garbuzov, personally pressured a number of other ministers to vote against the project. The situation was made even more difficult by an article in Izvestiya (Известия) paper which mockingly asked “Will the Kremlin be controlled by a punch card?”. All this led that OGAS was denied funding in 1970. Though various attempts were made to revive the idea, fairly little was achieved by the dissolution of the Soviet Union in 1991.
1980s
In the 1980s things became even worse. In the Soviet Union the same trends continued: the demand for computers grew, but the industry could not keep up and the general lack of computer hardware became ever more acute. On top of that, during this period personal computers really took off in the West. This was tragic for the Soviets, because personal computing really played into their weaknesses: general shortages of computer parts, and lack of grassroots hobbyist community who could push the industry forward.
A major breakthrough occurred during the Perestroika years, when in 1988 worker cooperatives (or small private enterprises) were allowed by the Soviet law. This has promoted a number of talented engineers to attempt to manufacture and sell computers themselves.
Since there was no such thing as copyright in the Soviet Union, these engineers could simply clone a computer from the West and sell them. The computer that was cloned the most was ZX Spectrum from the Sinclair corporation. Spectrums appeared in 1982 and quickly became popular due to their affordability – standard unit cost only £125. This low price-point was achieved by simplistic design and decision to use only the most simple and cheapest parts. These features made these computers especially suitable for cloning and for production in the Soviet Union, where more complex microchips were not available.
This and other posts in this series are only possible due to extensive reading and desk research I’ve carried out. Check out the Resources section to see some of the most interesting books and documentaries I’ve identified along the way.