Hacker News new | comments | show | ask | jobs | submit login
Pioneers of Soviet Computing [pdf] (sigcis.org)
91 points by heyts 250 days ago | hide | past | web | favorite | 21 comments



Those working on regular expression matching or related tasks should be well aware of some of the names here, especially Victor Glushkov. We still use Glushkov's NFA construction (which dates to 1961 or earlier) in our regular expression matcher (Hyperscan). It was interesting to hear about what a major figure Glushkov was beyond his work on automata, which was all I really knew about.


> By the late 1960s Lebedev, Glushkov,and their followers believed that Soviet scientists had accumulated a significant amount of experience in computer technology and had a considerable production potential. They wanted to collaborate with large Western European computer manufacturers in developing a fourth-generation machine before the Americans did. Lebedev’s political adversaries proposed a different option – to duplicate the American third-generation IBM-360 system, created several years earlier. Although no scientists of Lebedev’s caliber were among them, they were the political figures who had decision-making power. The Soviet government passed a resolution to develop a Unified System of Computers, reverse - engineered from the configurations of the IBM - 360.

The history there takes a turn and after 70s they just started mostly copying Western machines. It was interesting how they would justify it to themselves copying the product of the "decadent and failing Western capitalism" while they also had to attend parades and sing songs to Lenin about how their country was at the forefront and leading the world to better and brighter future.

However, to be fair, I also learned computers by using a crapy Soviet ZX Spectrum clone. So on a practical level copying became an obvious choice perhaps. At least it would have been soon enough.

Also, I always liked Setun, the ternary computer. It ends in kind of a sad story though:

> Unfortunately, after the Setun-70 project, Brusentsov’s lab was relocated from the Computer Center at Moscow University to a windowless attic in a student dormitory and was deprived of any serious support. The new university rector considered computer design a pseudo-science. Brusentsov’s original Setun computer, an experimental prototype that had faithfully worked for seventeen years, was barbarically destroyed and carted off to the dump.


Russian wikipedia https://ru.wikipedia.org/wiki/ЕС_ЭВМ#cite_note-.D0.B8.D1.81.... says that second generation soviet systems were oriented towards numerical computations and that they had a real problem with system software (no OS, even no assembler).

Now the architecture/os/system of the IBM S360 took some money to develop - here it says it was 5 billion $ (comparable in magnitude to the moon program - that was 25 billion $) http://www.smashcompany.com/business/cost-overruns-and-the-i... . So I guess that Kosygin was trying to avoid this expenditure.

Therefore a major point is that software development costs a lot, and they could just pirate that by copying the OS360.

On the other hand they could have done a similar thing as Amdahl/Siemens/Hitachi - copy the architecture while producing a hardware design of their own, so go figure.


They did do exactly that in many cases - the ES systems were not transistor-by-transistor copies of the IBM systems. The original article (PDF link) has a letter from Andrei Gagarin which mentions that the ES1065 had entirely original hardware design and architecture (while presumably being ISA-compatible with the IBM systems). Also, not all ES systems were IBM 360 compatible - the ES project was an overall scheme for improving and modernizing the computer industry of the Socialist block (the name ES was in fact created when Bulgaria, Czechoslovakia, Hungary and other countries joined the project, which grew out of a USSR-only project called "Ryad"), and although the IBM 360 architecture was central to ES, there were completely incompatible systems in the ES line.


Incidentally Lebedev didn't want to have anything to do with the cloning project (ES), I wonder if that was a mistake as he could have done a better job of preserving his project by doing a similar thing like Amdahl.

Actually that would have been better for the Soviet Union, as they would have been able to export a competing IBM compatible design for hard currency (opposed to just copying it down to every detail).

East Germany and other socialist countries were producing licensed IBM machines and they could no longer export these machines after the ES system got adopted.


> The history there takes a turn and after 70s they just started mostly copying Western machines. It was interesting how they would justify it to themselves copying the product of the "decadent and failing Western capitalism" while they also had to attend parades and sing songs to Lenin about how their country was at the forefront and leading the world to better and brighter future.

Oh, COME ON. Did you really think that every Soviet citizen was completely brainwashed? It was PROPAGANDA. It was not true by definition. Almost noone believed it. Your sentence is as disingineous in the same way as me wondering how Americans can build planes, vaccines and green energy plants when OBVIOUSLY they're all rednecks that shoot guns on their backyard and don't believe that global warming exists. Soviet people were significantly more aware of the shortcomings of the political system - after all they were the ones being the most affected by economical issues.

Remember that US propaganda from cold war wasn't significantly weaker when it came to USSR and a lot of things you heard repeated constantly are hugely exaggerated, wrong and built in times when authors of texts didn't really have insight into what's going on inside the country and had motivation to omit things to fit their narrative.


> It was PROPAGANDA. It was not true by definition. Almost noone believed it.

I think that you fail to entertain the possibility that a big part of the population did believe it. Propaganda might have been so prevalent because it worked. I remember seeing a documentary with some of the detainees in work camps in Syberia and they were saying thins like "Ooh, if Stalin would know what they are doing to us here."

As a inhabitant of an ex-communist country, I can tell you, the effects of propaganda are deeply rooted in the common consciousness and dismissing such things outright would be rash.

Perhaps highly intellectual people were more aware of their situation and their shortcoming as a nation, but even so, I am sure they had their vision clouded.


> Almost noone believed it.

As an ex-Soviet citizen - you're very wrong.

And for innovations vs copying - most of the Soviet industries were based on copies from "decadent West" (legal or illegal).


> It was interesting how they would justify it to themselves copying the product of the "decadent and failing Western capitalism" while they also had to attend parades and sing songs to Lenin about how their country was at the forefront and leading the world to better and brighter future.

A lot of (if not most, especially educated) people knew USSR is a piece of shit and were happy to bring in at least a wee bit of Western goods. My grandma studied & worked as engineer during USSR golden years and her favourite joke is "our ships are largest, our airplanes are largest, our microschemas are largest".


There was a joke in the USSR mocking the habit of bragging about everything Soviet being the best: "Наши микрокалькуляторы-самые большие микрокалькуляторы в мире!!!"—"our calculators are the largest in the world". Calculators being called "microcalculators" did not help.


> Setun, the ternary computer

The lab still exists actually. It was shuffled around and renamed, but it still has people that were involved with Setun-70.

https://cs.msu.ru/laboratories/203 (in Russian)


I'm pretty sure e is optimal encoding of wires. e is closer to 3 than 2, eventually everything will be ternary. Vindication may take a hundred years, or more, but that will absolutely be present in the computation endgame.


Kind've post-Soviet but there was a link between Elbrus (Boris Babayan) and Transmeta.

https://www.theregister.co.uk/1999/10/24/russian_merced_kill...

I believe that Babayan was the originator of the Code Morphing idea. However, Ditzel had also done something earlier with CRISP at Bell Labs where complex instructions were translated/decoded into horizontal microcode instructions and cached, the precursor to the x86 μ-store. Translated != Compiled.

https://www.technologyreview.com/s/400837/the-software-chip/

Babayan went on to work for Intel where he is now an Intel Fellow.

https://newsroom.intel.com/biography/boris-a-babayan/


> Elbrus, VLIW, Transmeta, Intel

It's interesting how much people Intel soaked up in the early 2000s after the RISCollapse and VLIW-gone.

Sometimes I wonder where exactly Intel got all the money from for doing that and pushing through quite some years with inferior products.


Read this years ago and loved it. Among the most memorable portions of the book for me is still the forward:

"Bringing this manuscript to publication was an epic adventure in itself, so we decided to share our experience here, along with some commentary on academic publishing today and its inevitable demise."

I've also quoted it before on Hacker News:

https://news.ycombinator.com/item?id=4059080



I don't see Valentin Turchin mentioned in the book which is a shame, he was arguably the Soviet counterpart to John McCarthy.


A claim like that is better with links that show it. Especially if compared to John McCarthy. Show us what you have.



Neat. Thanks. Main thing I wanted to look at is in Springer. Ill have to dig up non-paywall stuff on him later. Especially the meta and supercompilation.





Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | DMCA | Apply to YC | Contact

Search: