[throwaway_65123]

Fun fact -

The computer science department of the Moscow State University [1] was created in 1970 as a joint effort between Physics and Math departments to address computational needs of Tokamak models. Hence the official name - the department of computational mathematics and cybernetics.

The joke goes that they were hoping to "solve the Tokamak" in a couple of years and then move on to other fields of computer science - a dream that is still very much alive.

[1] https://cs.msu.ru/en

[ivan_gammel]

I guess it was not only the fusion that interested Soviet scientists. I remember attending lectures on parallel computing there, where Vladimir Voevodin was telling the story about Soviet computers in 1970s calculating ICBM flight routes 30 minutes faster than American ones. Not sure if this story was true, but there was always a significant military component in research.

[calamari4065]

>Soviet computers in 1970s calculating ICBM flight routes 30 minutes faster

That's hilarious. That sort of computation has to be completely trivial with modern computers. It's so crazy to think that plotting a ballistic trajectory ever took more than a few milliseconds, much less hours

[jazzyjackson]

IIRC the press unveiling of the ENIAC boasted that for the first time, they could compute a parabola in less time than it took for the projectile to fly it.

The next milestone years later would be "predict the weather 24 hours from now in fewer than 24 hours"

A lot of the government's interest in advancing computers was the military advantage of being able to predict the weather. Maybe if it was thoroughly understood, they could even control it! Alas.

All covered in the book "Turings Cathedral"

[namibj]

They are ((sub-)orbital) rockets, not atmospheric bullets.

I'm not surprised it was hard to compute to the accuracies desired.

[calamari4065]

The B in ICBM stands for ballistic.

I'm not saying the computation isn't complicated, I'm marveling at the fact that my watch can run that computation in microseconds

[ivan_gammel]

You are right, it’s complicated. Ballistic doesn’t mean it is a parabolic trajectory. At this distance (I stands for Intercontinental) you need to account for the atmospheric drag, winds, differences in gravity etc. Besides that ICBMs of 1960-1970s were using inertial navigation and had maneuvering warheads (e.g. SS-10), which increased complexity of the calculations. You still can do it on your modern watch.

[mustafa_pasi]

Interesting. Computational mathematics has really very little to do with traditional CS topics.

[thriftwy]

Traditional CS is not in vogue at ex-USSR universities. Poring over type systems and pure functions just never caught up over here, compared to good old Knuth-worth "close to metal" CS.

Some SDEs do take interest in these topics independently, though.

[DeathArrow]

What do they study in ex-USSR universities?

[ivan_gammel]

CM&C MSU masters -https://cs.msu.ru/sites/cmc/files/docs/sopostavlenie_magiste...

MIPT - https://mipt.ru/english/edu/master/

There’s everything, including traditional CS topics (compilers etc), AI and quantum informatics.

[avmich]

I sometimes feel sad about one particular branch of study, which could be applied to speed up computers, which was active right when the USSR was collapsing. Konstantin Likharev, among others, from Moscow State University worked on Josephson junction logical gates which demonstrated - then, 30 years ago - switching frequencies which are still out of reach for at least the mainstream CPUs (in some hundreds of gigahertz). A bit hard to find good articles about that research today - this one, https://www.beilstein-journals.org/bjnano/articles/8/269 , is an example. I've heard it failed to produce a commercial product because it was really hard to miniaturize.

[thriftwy]

Data structures, low-level workings of CPU and other hardware, computational compexities, binary algebra, combinatorics and discrete maths. Throw in some logic circuits design.

The same things they ask on FAANG interviews, coincidentally.

[flopriore]

It seems very similar to what we study in "Computer Science and Engineering" in Italy, which is a different degree from "Computer Science". The former is considered engineering, the latter isn't (pretty strange tbh, but that's how it works)

[closewith]

What are the traditional CS topics?

[specialist]

Sorting, searching, caching. (ducks)

[FiatLuxDave]

In case anyone missed the ducks in the traditional CS curriculum:

https://en.wikipedia.org/wiki/Rubber_duck_debugging

https://themindcollection.com/atwoods-duck/

[timthorn]

Naming things.

[specialist]

Ya, I second guessed myself. First in, then I thought "Naw, some pendant will say 'ackewlwally, that's software engineering'". Next time, I promise to take the principled stand.

[flashback2199]

CS not software engineering.

[biomcgary]

off by one

[belter]

Once and only once delivery

[belter]

Once and only once delivery