Soviet Computer Software and Applications in the 1980's [pdf] | Hacker News Reader

Soviet Computer Software and Applications in the 1980's [pdf]

ucis.pitt.edu

90 points_lnwk23 comments

It's a very interesting document. Must have been really boring to be a Soviet programmer, just cloning western stuff down to the exact commands all day.

One reason the paper is interesting is that it seems clear eyed. Popular press articles about the USSR's failed attempts to build an internet or computer industry always try to put an incredibly positive spin on it. The AEON article is titled "How the Soviets invented the internet" but if you read it, you'll discover they never invented anything. The story of the Soviet internet never gets beyond a bunch of dreamers having those dreams repeatedly squashed by the central planners who didn't care or actively wanted to undermine them. They never did any actual R&D, like by designing a TCP/IP equivalent, they didn't set up real functioning networks. They basically said to the government "wouldn't a computer network be cool?" and the government said "no not really". End of story. It's silly and misleading to describe this as "how the Soviets invented the internet".

There's a less detailed but more realistic take from the BBC here:

https://www.bbc.com/future/article/20161026-why-the-forgotte...

Even so it ends by saying "The Soviets may have lost the net race, but they were definitely in the game." even though by the BBC's own telling, behind the iron curtain the game never even started.

IC development was even worse. We (some of my professors) were reverse engineering western CPUs etc. for domestic consumption. This involved gently shaving top of the package and retracing the circuitry for lithography.

There were moments of genuine excitement however, when the sandpaper caught some of the circuit one had to imagine what should have been there. Thus creating novel design.

Did you have to do things like that because you were ordered to by the state, or because there was no funding to do any actual original design? Why did the industry never create anything new - presumably the Kremlin didn't actually say "don't be creative" to you and it was an emergent property of the system?

In socialists systems all incentives are aligned to gamble the system. You do gamble the system by fullfilling the bare minimum demanded from you and squeezing the most out of the rest of whats handed to you, because your material requirements are huge.

Most likely, most of the resources for computer sciences were traded away for others amenities before they ever reached a computation lab. The mentality ist still there in the old citizens of the eastern block. All that remained was the bare minima to do the chop required by quota: "Copy that chip"

Thanks! By the way, the term you're looking for is "game" not "gamble" i.e. people were gaming the system.

Ah, years of using it wrong come back to haunt me.. thanks

All three. The order came from above, that's how command economy (what socialist regimes generally implement) works. However, that order was result of insufficient funding, which left nothing to set up independent research pipeline in the area.

At the same time, Kremlin did often stifle development on purpose, under guise of "common good". Especially in dependent states across Europe and smaller Soviet Republics. In theory, this was meant to prevent internal competition and waste of resources. In practice it was a way to sell inferior goods and create dependency on the central authority. This happened with everything, computers, trains, cars.

So, a very imperialist Kremlin.

I'd say that reverse engineering an 80's CPU was everything but boring.

DEC's first VAX microprocessor was cloned¹. On the successor CVAX² die, DEC included Russian text³ saying (supposedly) “When you care enough to steal the best”.

¹ https://en.wikipedia.org/wiki/MicroVAX_78032#Other_vendors

² https://en.wikipedia.org/wiki/CVAX

³ https://i.stack.imgur.com/xmYaT.jpg

FYI, that text is completely incomprehensible in Russian. It looks like the chip designer translated an English phrase word-for-word using a dictionary, didn't conjugate verbs or decline adjectives, added some gratuitous misspellings, and ended up with Engrish type nonsense.

I think the situation had vastly improved by the mid-late 80s. Contrary to the prediction in the paper in the original post, by the 90s there was a burgeoning software scene in Russia (and I'm sure other Soviet states). My dad worked at a manufacturer where, in the late 80s, he developed FEA software that he still sells to this day (https://quickfield.com).

A former co-worker of mine started a CAD/CAM software company in Russia in the late 80s-early 90s. He was targeting soviet x86 & VAX clones. Business was not bad; but the weird postsoviet economy made it untenable.

He told me:

"I was making good money on paper; but still, I had to spend a whole month's paycheck to buy four new tires for my car. They were not even particularly good tires, and it wasn't a very good car!"

Central waterfall-like planning for software at the company level is harmful.

Imagine this at the national level.

The soviets (according to TFA) were depreciating their CAD/CAM setups over 20 years. So Госплан probably thought they weren't doing waterfall — after all, they had short, 5 year, sprints (пятилетка)!

Many programmers were busy doing payroll processing software for the various organization. Each org had to have it's own thing for its payroll (which is sort of surprising)

It really fast converged to using Excel by accountants themselves or using/programming "ERP" 1C[1].

[1] https://1c-dn.com/

Maybe there was something to be said for "digital literacy" after all; I used to think it a useless term, but skimming this report, observing a relative desert, makes me think that I grew up as a fish in water:

According to one recent source, Soviet "outlays for software did not exceed 1.5 to 2 percent of overall outlays for computer hardware."

Glushkov, a founding father of Soviet computer science and a network visionary without peer, wrote his 1979 history of Soviet computerdom without a single mention of networks.

Designers with CAD work stations faced the choice of using generally available but poorly supported subroutines or of developing their own software.

Note especially: CAD, to the extent that it fulfills its promise to optimize the use of materials, reduces the cost of the project and thus also the payment to the design organization which, therefore, has little incentive to use it. !

While the hardware requirements have slowed CAM development in the Soviet Union, the necessary organizational changes, from new job descriptions to a firm's relationship with its suppliers, are perhaps the largest barriers the Soviets will have to surmount for the successful implementation of CAM.

As of early 1987, the Soviets had not solved the problem of providing maintenance and support service to enterprises with CNC machine tools and other advanced manufacturing technology.

When, for instance, the USSR Committee of People's Control conducted a random check of robotization efficiency three years ago, the picture that was revealed staggered even the inspectors who had seen a thing or two before. The economic efficiency of implementation of 600 robots that cost over R10 million was equal to . . . R18,000 a year. And the chance to free up one worker was paid for by ... 14 robots on average!

One final finding of the Leningrad area survey merits mention. In 1983, on average, the computers of the area were out of commission 9.3 percent of a working day, of which more than half was due to technical breakdowns.

It's interesting to note that in the west, Computer Aided Manufacturing LONG predated Computer Aided Design.

"A Possible First Use of CAM/CAD" is a good read.

https://hal.inria.fr/hal-01526813/document

Abstract: This paper is a discussion of the early days of CAM-CAD at the Boeing Company, covering the period approximately 1956 to 1965. This period saw probably the first successful industrial application of ideas that were gaining ground during the very early days of the computing era. Although the primary goal of the CAD activity was to find better ways of building the 727 airplane, this activity led quickly to the more general area of computer graphics, leading eventually to today’s picture-dominated use of computers.

Thanks, buildjets, and thanks, Boeing hackers:

This made the 1103A the world’s largest, most expensive punched tape machine.[0]

[Computers] could not draw lines on paper accurately enough for design purposes. ... We simply replaced the cutter head of the milling machine with a tiny diamond scribe ... and drew lines on sheets of aluminium.

Also: ...the somewhat loose airframe manufacturers’ computer association got together to write an APT compiler for the IBM 7090 computers in 1961. Each company sent a single programmer to Convair in San Diego... It sounds as if the airframe manufacturer's computer association had better, or at least more effectively cooperative, "soviets"/"workers' circles" than the ОКБs.

[0] for reference, the 1103A had a clock speed in the kHz, and weighed ~17'500 kg.

Edit: Postscript — is the "Professor Doug Ross" mentioned in these documents the same "Douglas T. Ross" mentioned in the acknowledgements of Ivan Sutherland's thesis? (Sketchpad 63)

PPS: Yes- https://conservancy.umn.edu/bitstream/handle/11299/107611/oh... (DOD 1, ВПК 0)

Anyone have the companion papers?

Soviet Computers in the 1980s: A Review of the Hardware

Is one. I can’t find a copy not behind a pay wall.

Hug of death? :(

Works for me, but it's also on archive: https://web.archive.org/web/20220815211059/https://www.ucis....

Reply on news.ycombinator.com