We publish the continuation of the interview with Alexei Safronov , a researcher of digitalization in the USSR (for the first part read the link ) In the second part, Alexey tells what technical difficulties the ASR had in 1960 - x, that this one was able to the system before it was destroyed, and why it was in some ways more advanced than the current developments and projects of Russia in the field of digitalization.
Safronov. We only have three people left from Soviet times, all three are technicians. They said that at the end 30 - x there was a constant problem: Soviet computers were very unreliable. Kossov Vladimir Viktorovich, one of the former leaders of the Center, told me that the average task of the interindustry balance with the detail that was then being done was solved 40 minutes, and the average uptime was minutes. And regularly it turned out that the car broke down, not having time to count. In addition, the technicians constantly had mutual complaints with the operators. When something broke, they said that it was because the operators did not withstand the parameters of temperature, humidity or dust in the building. They answered them: “Everything is working fine for us, you don’t know how to monitor computers.”
Very often, when I asked for these stories, everything rested on the fact that the material base was far behind. It turned out that the initiative is punishable. Imagine that in order to establish remote access, you need not only to draw a grid, you also need to come up with a computer, find people who will make it, and also make sure that they do as they should. At one time, some of the people from the computer center received a state award for the creation of the Iskra - minicomputer) ", A clone of the American computer Wang . It was one of the first personal computers, with its help planners from the State Planning Commission could connect to the information base in the MCC. One of the main ideologists of the ASPR system, Vladimir Borisovich Bezrukov, said that in addition to the absence of some components that had to be developed independently like that, they were constantly brought down by technical reliability - the existing equipment often broke or did not work very well.
XX 2 CENTURY. 87611 If the technical level is in the middle 30 - x would be like in the middle 40 - x, could the project be brought to its logical end? And maybe then it would be possible to get out of the crisis?
Safronov. Now this is a frequent argument that computers are ripe and socialism is technically possible, but I'm a skeptic in this sense. If we read some kind of perestroika journalism, we see that the first market experiments began immediately after Brezhnev’s death, back in Andropov’s times, in 1983 year. And although the thesis that “we want to add a market” does not sound directly in this literature, it speaks of “development of independence”, of full cost accounting. Euphemisms of the market were used. The argument itself was far from the fact that "we are technically unable to cope with the plan." The thesis that "our computing power is insufficient so that we can normally calculate the parameters of the economy" did not sound. They said: “Our employees are not interested in working well, because there is no connection between the quality of their work and their remuneration.”
I do not think that the problems were on the side of technology. Because even if remote access didn’t work very well, this problem was solved. For security or technical reasons, if the planner from the ASPR data bank needed something, they would bring it to him in printed form or on a reel. There was a staff of people who ran with packed envelopes and delivered information on printouts from the MCC to planners upon request. (GVC and Gosplan were in different buildings in different areas of Moscow). There was a staff of typists who were supposed to drive these printouts back into the computer. Moreover, the principle of “two keys” was implemented there: to accidentally prevent a typo, two typists drove the value at the same time, and it went to the database only if both entered the same thing. After that, the planners uploaded their results back in machine form. It turned out that from Gosplan to the Main Computing Center the information was transmitted in a machine form, and from the Main Computing Center to Gosplan the necessary extracts were issued in paper form. This is not very convenient, but in any case this is not a problem that cannot be solved.
If the Union had not collapsed, then all these technical roughnesses of the system would sooner or later be debugged. It was planned to finish it completely by 1990 year. It would make it easier to solve many of the problems that we face today. Even now, a lot of effort is being spent on reconciling various forms with each other. For example, when the Ministry of Transport is building something, and the Ministry of Finance is financing this construction, an incredible amount of time is spent on comparing numbers from one form with numbers in another form.
XX 2 CENTURY. If you look at what is now, by what parameters can you compare the modern digital control system and the past?
Safronov. Currently, work is underway on the concept of the National Data Management System (NSMS). It is needed for this: we have many departmental systems, and these systems store a lot of data that they have accumulated for their needs. But for third-party organizations, this data, as a rule, is only partially available. There is an idea that the NSDS will allow us to ask the question “Do you have such and such data” not to each agency separately, but that there will be a single register in which all these departmental data will be sent. At the same time, it is not planned to replace departmental systems with one centralized system, but it is planned to create a resource that will know that they have this data there. The idea is that we create a registry that knows about all the existing data that is now difficult to find.
The implementation of such a system will also help solve the encoding problem. The coding problem arose when creating the ASPR. Any economic information needs to be presented in a uniform manner. I think everyone has come across this one way or another. For example, in the table of names of employees, you can first write a name, and then a surname, or first a surname, and then a name. If we have two tables differently formatted, then the computer will not understand that we are talking about the same person. Now there are smart systems that somehow catch these “fleas” themselves, and then it was possible to deal with this only by issuing some standard valid for the whole country, for example, how to record a cargo turnover indicator. And there were thousands of indicators! In addition to the task of developing a standard, it was necessary to force everyone to use it. Because of this, the MCC was forced to engage in norm-setting activities and prepare draft legal acts in the field of digitalization.
XX 2 CENTURY. 2020 Is the ACM similar to what Starovsky suggested?
Safronov. Such a thing - one place that knows about all state data in all industry systems - would be in demand both by the State Planning Commission and the CSB. Everyone needed data to work. But there is no planning provided in the NSDS, there is no optimization model, so this is definitely not the scope of the OGAS, even in the truncated Gosplan version of the ASPR. Actually, I didn’t think about it that way, but the NSUD project really is similar to Starovsky’s version. It turns out that we had a maximalist version of cybernetics enthusiasts, a moderate version of the State Planning Commission and a minimum version of the CSB, and now we are trying to make the minimum of the three approaches that appeared in the middle 1960) s. By 2019 the year came to attempts to make the minimum version of the three, falling apart before the moderate, which was already somehow done, albeit at very ancient elemental base.
And even this completely truncated version of the system is implemented with great difficulty, because data is money. For example, in Russia a few years ago, the Plato system appeared, in the framework of which money is taken from trucks for breaking roads. For this, all trucks are equipped with Glonass; this system sees all freight traffic on federal highways. This information can be very useful for the formation of an inter-regional transport balance, from which regions to which products are transported. Suppose if we want to build a new road, for this we need to understand whether they will drive along it or not, and if so, which regions will get the most effect from this. When the Ministry of Transport first applied for this data, Plato said: “Why should I give you something?” The Ministry of Transport replies: “I gave birth to you!” "Plato" objects: "And you signed my charter? “Open it, look, is there something written that I have to provide data to someone?” Then we agreed, of course, but it took a long time.
Any ideas of large information systems or associations of information systems are greatly hindered by group interests, namely that money can be made from data. If this is logistics data, then it can be sold to logistics companies that need to build cargo flows. Let's say we have a navigator from Yandex, in which we can build a path, but if we need the same thing for trucks? To do this, we need to know that some kind of bridge does not withstand so many tons, but somewhere under the bridge a car is passing by, but there is no truck. Or that somewhere the road is bad and it is better not to ride it again. If you know these additional things, you will carry faster, and if so, then you will lure customers from competitors. On the other hand, if someone else knows my data, then he may understand that I am not working well. There is a fear that “we will tell you everything about ourselves, and then you will poke us into this, say, we know from your data that you can work better.”
XX 2 CENTURY. 87611 But does the System of Interagency Electronic Interaction (SMEV) not solve the problem of data centralization?
Safronov. Partially. There is tangible progress with the number of public services provided electronically, departments have learned to exchange information so that citizens do not need to manually fill out the same questionnaires ten times. But in general, access to data and the exchange of data (rather than letters) is carried out only where people are legally obligated to do so. “Voluntarily” no one will open their data to the world.
"Beyond" public services, electronic document management now works as follows. The official opens an internal database and writes a letter with some information from there. This draft letter is signed by its leader - previously manually, now more and more by electronic signature. If this letter is, then it is not so scary, but if it is a report on pages, then on the other side there should be a person who opens the PDF and either recognizes it as a file reader, or he finds the digit from it and hands it to the Excel file. You can exchange documents, or data, and now the electronic document management system (EDMS) is basically an exchange of documents - scanned papers. But this is not data exchange! Data exchange is when we have access to the source. There are advances in this business, but the process is much slower than we would like.
XX 2 CENTURY. The program of introducing e-government technologies "Electronic Russia" was launched quite early in 1994 - x. In 2000 - m it ended with almost zero results. Those who were engaged in it simply did not know about the existence of relevant developments in the USSR?
Safronov. I think so. As a decent person, I searched in 2000 - 2018 for his master’s years any articles on ASPR and could not find modern articles. There are mentions in memoirs, there are many general proclamation articles that would be worth using the experience of the State Planning Commission. Specifically, the knowledge that such a system existed, it seems to me, has been forgotten, that's all. It in some form, on the old stocks, continued to work until the beginning 1990, while Yakov Moiseevich Urinson worked in the MCC. There were intersectoral balances built on the basis of Soviet interconnections. The system was already used only for forecasting, but not in order to make plans, because there was no one to give orders for implementation. Then Urinson was taken to the Ministry of Economy, and after that traces of the ASPR are lost.
We still had one of the computer technicians at the MCC who retired last year, and I asked him where everything has gone. He said that in 1983 - s brought new computers, already IBM , and the entire range of ASPR programs and a data bank stored on an EU computer, simply did not transfer to new media. All this remained on those sunken cars.
This story is about how years to make an economic management system, and then just say: "We have an old hard drive here, and let's throw it out."