Comment by mbarti | Hacker News Reader

Comment by mbarti | Hacker News Reader

Parent Full thread

So the Polish intelligence selected 3 smartest mathematicians in 1929 and assigned them to the task of understanding the German encryptions. The team (Rejewski) had a breakthrough in 1932 and managed to break the Enigma codes. By 1939 they've broken many subsequent versions of the constantly upgraded ciphers.

In 1939 they gave the British: a working Enigma machine, the broken encryption, a full description of how the encryption works, how to use a machine to speed up the process (the "bomba" appliance), and what automation is needed for the following versions.

The British (Turing et al) then contributed a machine-based automation which allowed to break the following upgrades to the Enigma ciphers, which would be impractical to manual calculations.

As a side note, the Polish codebreaking bureau started with a breakthrough around 1917, when some smart soldiers managed to break Bolshevik radio communication ciphers, which allowed Poland to stop the Russian revolution's expansion by 1920 (the decisive Battle of Warsaw).

There is a small memorial (to the Poles) at Bletchley Park. I went there on purpose once to see it, and unfortunately I didn't have time to go find it again when I was there earlier this year, but it's a memorial so they won't have removed it.

The (recreation of the) machine for breaking wartime Enigma is now in the Computer History museum, rather than the main Bletchley Park museum (they're on the same site but open on different days - check before travelling!), which is slightly incongruous as it's just not a computer. The museum also has several machines which are each arguably the first computer or an early computer (it depends on what in your opinion distinguishes a computer and how hard you're willing to squint at your definition) including (again a reconstruction of) Colossus, which was built for Bletchley to attack the actual interesting cryptography of the war, Lorenz - not Engima and W.I.T.C.H (the Harwell computer) but the choice to put the reconstructed Bombe in that museum is a bit weird.

I was just there. Easy train ride from London. Enthusiasts can spend 6 hours, or more there. Not a great family destination, though. Sometimes you have to wait hours for your tour start. During that time you've burned all the good will with the family as you've briskly seen all the stuff the tour will now show you.

I will go back on my own one day.

Ha, yes this was my experience exactly!

There is also the national radio center to visit which I didn't get a chance to visit unfortunately.

http://www.nationalradiocentre.com/

And the National Museum of Computing is there as well.

https://www.tnmoc.org/

Polish memorial was in Bletchley Park when I visited in 2019 under some trees:

https://maps.app.goo.gl/1Mot8nCCvS3e6r9SA?g_st=ic

"This is where the first break of Enigma by an exclusively British team took place."

https://www.google.com/maps/@51.9971405,-0.7429412,2a,25.5y,...

Lol, some memorial.

I think it can still be considered a computer, just not a general purpose programmable computer right? I'd expect adding machines and stuff to be in a computer history museum.

It's even more primitive than, say, a mechanical calculator of that era, which yes, the NMC does have in a cabinet near the start. The Bombe isn't doing even simple arithmetic. It's essentially an electric motor, plus a lot of very complicated electrical wiring, to verify whether some particular setup (say, rotors I, V and IV in that order, set to JOM initially) could or could not ever cause the crib (a guess at some text an operator intended to send e.g. "Heil Hitler") to be converted into the ciphertext being attacked.

Each such setup is tried, very briefly, and if it couldn't be correct, the motor winds to the next, and the next and so on, until one setup which could be "Heil Hitler" is found. This is a "stop" - the machine literally stops until set going again. Women running the machine note down the exact setup which stopped it, and elsewhere very smart people try assuming this stop is correct - it might be, or, maybe the machine will stop again later, or, maybe the crib was wrong, maybe this message has a typo, "Heil Hiter"...

It's fantastically complicated but only in the same way that for example the movement of swallows is complicated. There's a relatively simple rule, it's just followed to an extreme degree and so that's impressive. It's not really very clever, and it's nothing like a computer.

Colossus isn't a programmable computer. It's for attacking Lorenz, so the people who built it didn't care that to attack a slightly different cipher the machine must be taken to pieces and re-assembled accordingly, they were here to attack Lorenz. In the rest of the early gallery there are several machines which are, or aren't, depending on how you squint truly programmable, truly electronics, truly storing the program and so on. But the Bombe is to these machines as the Super Bowl is to tennis. It's not that it's a very primitive and strange game of tennis, it's not tennis at all.

Yes. Those machines were key-testers, not general purpose computers. Their modern descendants are Bitcoin miners with custom ASICs. They do one thing, very fast.

As with Bitcoin miners, what led to success was mass production. Bombe-type machines were built in quantity by the British Tabulating Machine Company in the UK, and the National Cash Register Company in the US. There was also a Western Electric prototype built from telephone relays.

Interestingly, the US cryptanalysis effort against Japan was statistical, rather than brute-force search.[1] Much statistical work was done using IBM tabulating machines.

[1] https://web.archive.org/web/20210828044031/https://cryptocel...

It isn't programmable but neither is an early adding machine. I don't see a problem with either being in a computer museum.

It's unfortunate that in pop culture, Turing gets all the credit. Even the fairly recent movie doesn't do the justice.

Turing wasn't a publicly known figure for many years until a campaign to give him some prominence and highlight the injustices done to him started to get some ground. There was a strong gay rights thread to this, which obviously didn't apply to others.

While I agree others should no doubt get credit (Turing's coworkers also), the Computer History Museum and Bletchley Park do a good job (last time I visited) at describing everything relatively impartially. Media (screen and print) could often do a better job though.

I’m quite interested in the subject, but literally never heard of Rejewski until now and it feels pretty terrible. A lot gets said about other biases, but Poles seem to not get a lot of credit pretty consistently and often get appropriated by the host countries they had to work for. Hope that’s all in the past now…

Because cold war propaganda, and even some propaganda issues during the war.

Bad enough that despite half a million army on the western front, it was effectively banned to take part in London victory parade.

Have a read of the Charge at Krojanty for an example of Nazi propaganda. A myth that perpetuates till very recent.

The Imitation Game doesn't do justice to Turing himself, either. Pretty dreadful all around.

Pretty bold to write Turing as a socially inept autist loner with a temper problem when there’s zero historical evidence for any part of that characterization, only evidence suggesting the opposite.

And IIRC with a female romantic interest... that to me was the worst offense. It was an awful movie all around.

I've not seen the movie, but Turing was engaged briefly to Joan Clark, a fellow cryptanalyst, and they were close friends. She knew about his homosexuality though so it would have been literally a marriage of convenience, probably not uncommon at that time.

As I recall it was pretty clear what he was actually interested in, and it wasn't her, so it at least wasn't completely wrong on that particular point. I think it was explicitly sort of a show/cover relationship. But of course I saw it once a long time ago...

I suspect it was pretty good if you don't know or care about Turing. But yeah, it pissed me off pretty bad.

It is disgraceful that they used Turing's name in that film, was was almost completely fiction. Everyone involved in it should feel ashamed.

Interested viewers may enjoy a somewhat more historical treatment in the BBC's "Breaking the Code", a 1996 made-for-TV movie starring Derek Jacobi.

https://www.imdb.com/title/tt0115749/ https://www.youtube.com/watch?v=udW0j96vAOk

This movie was terrible and I thought hugely offensive to the memory of Turing and all involved. Awful awful awful.

Don't even bring that movie into this discussion, it's a historical atrocity.

> So the Polish intelligence selected 3 smartest mathematicians

The interwar period in general saw an extraordinary flowering of top notch mathematicians and mathematical logicians in Poland [0][1], including Tarski and Banach. So the conditions in the country were very conducive. Others besides Rejewski, Rozycki, and Zygalski also worked on deciphering enemy messages years before WWII, such as Lesniewski, Mazurkiewicz, and Sierpinski [2].

[0] https://en.m.wikipedia.org/wiki/Polish_School_of_Mathematics

[1] https://plato.stanford.edu/entries/lvov-warsaw/

[2] https://en.m.wikipedia.org/wiki/Jan_Kowalewski

Little more precize about Soviet cipher braking. It was a group of mathematicians, not just smart soldiers.

https://en.m.wikipedia.org/wiki/Polish%E2%80%93Soviet_War#Po...

Was the automation the more difficult of the two problems?

Why did the Polish mathematicians not undertake that step themselves between 1932 and 1939?

AFAIK the Germans kept improving both the Enigma itself and the procedures for using it. For example there was some specific weakness that the Polish attacks used, but at some point Germans figured it out and "patched" it. By the end of the war they could change the "encryption key" even more often than once a day, whereas before the war they could reuse the same key for many months. The Poles didn't break all Enigma setups till 1939 because one didn't have to it yet to effectively collect intelligence. Also because they hadn't existed yet :)

Enigma was actually pretty good, but it had a major design flaw: insufficient diffusion. The YouTube channel Computerphile made a good video on practical cryptanalysis of the Enigma using modern hardware: https://youtu.be/RzWB5jL5RX0. In short, if you know what you're doing you can do ciphertext-only attacks given long enough messages and minimal knowledge about the distribution of the plaintext (their example relies only on the fact that the plaintext is in English).

That's what the Bomba [0][1] was about. The Brits worked off those developments and continued the work that the Poles had started but were unable to continue themselves because of the "deteriorating political situation".

[0] https://en.wikipedia.org/wiki/Bomba_(cryptography)

[1] https://www.bcs.org/articles-opinion-and-research/the-enigma...

Reply on news.ycombinator.com