The crooked hand of Death itself
December 26, 2024 5:24 PM Subscribe
I never fully understood how an Enigma Machine works until I watched this. I think it could be appreciated just for the skill of its 3D animation, even if you’re not interested in World War II cryptography. [NOTE: The link begins the video in the middle after less interesting introductory material. This is intentional.]
OK this was a genuinely well-done video, way better than Planck Length was. Now please find me an equally good video that shows me how the Allies broke it. I'd love to watch that.
posted by outgrown_hobnail at 8:15 PM on December 26 [6 favorites]
posted by outgrown_hobnail at 8:15 PM on December 26 [6 favorites]
Enigma was available commercially starting in the early 1920s. Breaking the cipher happened over a long period of time and started well before the war: Britain tried and failed in the late 1920s, Poland succeeded in the 1930s but was later thwarted when Germany increased the rotor count, then Britain succeeded during the war after Poland shared its work on the prior version of Enigma. Wikipedia has an excellent summary.
posted by migurski at 8:37 PM on December 26 [8 favorites]
posted by migurski at 8:37 PM on December 26 [8 favorites]
That's the best Enigma explained I've ever seen, thanks!
posted by sotonohito at 8:47 PM on December 26 [1 favorite]
posted by sotonohito at 8:47 PM on December 26 [1 favorite]
Very well done for the mechanics, but I still don’t get the logic. Given the way a given setup keeps nudging the encoding rotors forward, I don’t understand how typing the encoded output into the receiving machine decrypts the message. It seems like it would encode it another 7-9 times from the original. How is the reverse circuit for the substitution constructed when the rotors only ratchet in one direction?
posted by bendybendy at 9:08 PM on December 26 [2 favorites]
posted by bendybendy at 9:08 PM on December 26 [2 favorites]
Very well done for the mechanics, but I still don’t get the logic. Given the way a given setup keeps nudging the encoding rotors forward, I don’t understand how typing the encoded output into the receiving machine decrypts the message. It seems like it would encode it another 7-9 times from the original. How is the reverse circuit for the substitution constructed when the rotors only ratchet in one direction?
I am not a cryptographer or engineer or electrician, but I assume that you change the battery direction so that current runs through the circuits in the other direction. You want the rotors to rotate in the same direction (which they will because that part is mechanical not electrical) so that the next letter gives you the same circuit as the original encoding.
posted by If only I had a penguin... at 9:15 PM on December 26 [1 favorite]
I am not a cryptographer or engineer or electrician, but I assume that you change the battery direction so that current runs through the circuits in the other direction. You want the rotors to rotate in the same direction (which they will because that part is mechanical not electrical) so that the next letter gives you the same circuit as the original encoding.
posted by If only I had a penguin... at 9:15 PM on December 26 [1 favorite]
Regarding decryption: I think the key is the reflector. Every letter is tied to another letter. J lights up B, B lights up J.
As the previous commenter mentioned, the advancing rotors create the same circuit each time.
posted by just.good.enough at 9:39 PM on December 26 [5 favorites]
As the previous commenter mentioned, the advancing rotors create the same circuit each time.
posted by just.good.enough at 9:39 PM on December 26 [5 favorites]
No wait...would the battery have to be turned or at any given setting are the letters reciprocal? Like if right now A would come out as T would (at that same setting) T come out as A?
posted by If only I had a penguin... at 9:40 PM on December 26
posted by If only I had a penguin... at 9:40 PM on December 26
Ah, ok, thank you...playing with my son's SnapCircuits has not taught me as much as I thought. :)
posted by If only I had a penguin... at 9:41 PM on December 26 [1 favorite]
posted by If only I had a penguin... at 9:41 PM on December 26 [1 favorite]
Yeah, the scrambling of circuits and the ratcheting of wheels is meant to make the next letter unpredictable. The reflector ensures that in any given state it's symmetrical.
The part I never understood is the plugboard. It seems like an almost superfluous addition, but apparently was key in making it so difficult to crack.
posted by ChurchHatesTucker at 9:49 PM on December 26 [3 favorites]
The part I never understood is the plugboard. It seems like an almost superfluous addition, but apparently was key in making it so difficult to crack.
posted by ChurchHatesTucker at 9:49 PM on December 26 [3 favorites]
From the movie, one of the keys to eventually cracking is was that there was repeated content across messages, so they could often guess that the last phrase in a message would be "Heil Hitler" for example. If you knew that was the last phrase and you knew the last letter in the message was a W for example, there are probably a relatively small (like let's say 3 digits, tops) rotor settings that take you from R to W. And then you know the second last letter in the message is an F. Then there are a limited number that take you from E -> F and of those only some could plausibly be one rotor rotation away from one of the R->W possibilities. The wiring within each rotor is fixed, so I could see how maybe if it were just the rotors and you could guess what one part of the message said, that you might be able to work out which rotors were used and in what position they were by the end of the message, which doesn't tel you where the ratchet catches are, but is still pretty good.
Now mess with that by adding the plug board which has no fixed positions and could be anything and everything above is kind of useless.
posted by If only I had a penguin... at 9:58 PM on December 26 [3 favorites]
Now mess with that by adding the plug board which has no fixed positions and could be anything and everything above is kind of useless.
posted by If only I had a penguin... at 9:58 PM on December 26 [3 favorites]
The Numberphile video is another good one I think.
posted by kiblinger at 10:54 PM on December 26 [2 favorites]
posted by kiblinger at 10:54 PM on December 26 [2 favorites]
Here is Numberphile on the fllaw that Turing expoloited to crack the code.
posted by If only I had a penguin... at 11:29 PM on December 26 [2 favorites]
posted by If only I had a penguin... at 11:29 PM on December 26 [2 favorites]
Yeah my understanding is the reflector creates pairs. So to decode, you hit the letter in the message which given the settings of the wheels etc, has been paired with another letter. Hitting if they are F and J, at those exact settings hitting one lights the other and vice versa. So by decrypting you can use that knowledge to unscramble.
posted by Carillon at 1:01 AM on December 27 [1 favorite]
posted by Carillon at 1:01 AM on December 27 [1 favorite]
From the movie, one of the keys to eventually cracking is was that there was repeated content across messages,
And from memory of, we'll say a youtube video or 2, another helpful element was a handful of Germans who never changed the wheel settings or bumped a setting by 1 every day.
The known setting users would also include the same signoffs to help verify the encryption settings.
posted by rough ashlar at 6:06 AM on December 27 [1 favorite]
And from memory of, we'll say a youtube video or 2, another helpful element was a handful of Germans who never changed the wheel settings or bumped a setting by 1 every day.
The known setting users would also include the same signoffs to help verify the encryption settings.
posted by rough ashlar at 6:06 AM on December 27 [1 favorite]
Another critical element to cracking it was the problem of initialization.
At the beginning of the day the tank commander hasgot the rotors set, the plugboard set, and they get a 57 character long message from regional HQ.
The problem is, regional HQ sent three other messages before sending the message to the tank commander, and so their rotors have already been incremented a couple hundred times and the tank commander's are in the starting position.
You face two problems here: either you reset the rotors to the initial position each message, which weakens your encryption becuse it's the same thing every message, OR you somehow send the number of rotor increments or the current rotor positions.
And of course sending your current rotor positions isn't exactly a great idea since it defeats the whole point of having the rotor positions set randomly.
There were various solutions, but all introduced new attack vectors for cryptanalysis.
Another thing worth mentioning is that Engima, while sophistocated, was also known to have vulnerabilities and even some people in German high command didn't really trust it. But organizational inertia, a misplaced trust in the machines and thier really nifty super duper complex setup by people who didn't understand cryptography, and the lack of a better replacement mean they kept using it.
And of course the Allies did their best to convince the Axis that Enigma remeained uncracked so they kept using it right up to the last.
posted by sotonohito at 6:26 AM on December 27 [3 favorites]
At the beginning of the day the tank commander hasgot the rotors set, the plugboard set, and they get a 57 character long message from regional HQ.
The problem is, regional HQ sent three other messages before sending the message to the tank commander, and so their rotors have already been incremented a couple hundred times and the tank commander's are in the starting position.
You face two problems here: either you reset the rotors to the initial position each message, which weakens your encryption becuse it's the same thing every message, OR you somehow send the number of rotor increments or the current rotor positions.
And of course sending your current rotor positions isn't exactly a great idea since it defeats the whole point of having the rotor positions set randomly.
There were various solutions, but all introduced new attack vectors for cryptanalysis.
Another thing worth mentioning is that Engima, while sophistocated, was also known to have vulnerabilities and even some people in German high command didn't really trust it. But organizational inertia, a misplaced trust in the machines and thier really nifty super duper complex setup by people who didn't understand cryptography, and the lack of a better replacement mean they kept using it.
And of course the Allies did their best to convince the Axis that Enigma remeained uncracked so they kept using it right up to the last.
posted by sotonohito at 6:26 AM on December 27 [3 favorites]
the Allies did their best to convince the Axis that Enigma remeained uncracked
The Bletchley Park work was not publicly acknowledged until 1974.
posted by Lemkin at 6:32 AM on December 27
The Bletchley Park work was not publicly acknowledged until 1974.
posted by Lemkin at 6:32 AM on December 27
OK, now here's the part I still don't get. What was the mechanism that prevented a letter from being coded as itself? Like I can see from the circuit in the video that a circuit can't go through a letter as input key and as lightbulb-illuminating circuit at the same time. But what keeps A letter A from coming out as a letter A? Like shouldn't it happen sometimes that an A comes out as an A and i can see why the lightbulb wouldn't light if that happened, but not why that encoding wouldn't happen. Like whatever letter the A comes out of the rotors as it's always possible that letter is connected to the A in the plugboard isn't it? Then what happens? I don't think I'm explaining this clearly. I can see why a letter encoding as itself wouldn't light a bulb but not why it wouldn't "try." Is it just that if no bulb comes on you try again once the rotors have moved?
posted by If only I had a penguin... at 6:57 AM on December 27
posted by If only I had a penguin... at 6:57 AM on December 27
So I thought my son would be really interested in this because he has been playing with SnapCircuits. He's 7. I showed him the video this morning and he thought it was really cool and that the Enigma machine would be very useful even today because "sure we have computers, but could computers possibly be as effective as that??" but then concluded that he thinks really the real solution to problems that the machine addresses is to find an eliminate the causes of war.
posted by If only I had a penguin... at 7:02 AM on December 27 [1 favorite]
posted by If only I had a penguin... at 7:02 AM on December 27 [1 favorite]
In a sense the issue is the three copper contact switches. If the switch is down to start the circuit at letter A, then it can’t simultaneously be up to complete the circuit and light letter A. I’m sure there’s lots of other stuff about how the cross connected rotors and plus worked, but that was the part I could get my head around in terms of why it’s non-trivial to fix.
posted by meinvt at 8:37 AM on December 27 [1 favorite]
posted by meinvt at 8:37 AM on December 27 [1 favorite]
This video on how a modern computer would be used to attack enigma is good.
posted by Mitheral at 8:59 AM on December 27 [1 favorite]
posted by Mitheral at 8:59 AM on December 27 [1 favorite]
Meinvut...yes that's the part that I understood. Why the lightbulb cant light if the same letter is input and output. The circuit wont complete. But i guess indont get why it would never happen that the configuration of the machine would "try" that combination. The circuit would never close because it would be open in that thingie under the key but what's to stop a setup where the route of the wires goes from A back to A.
Like inwoukd be curious to see a video where they run it with nothing in the plugboard, type an A. Let's say a T comes out. Ok now set the rotors bavk the original position and put in a plug from T to A... ah wait...there it is (at least part of it) ok so since the circuit goes through the plugboard first when you put in an A it would now become a T before going through the rotors so it wont come back to the T next time on that setting. It will go somewhere else.
But ok, what prevents that T from coming back as a T which then becomes an A... I mean the original problem again. Still unsolved, just that my idea for testing "what happens if..." doesn't work.
posted by If only I had a penguin... at 9:26 AM on December 27 [1 favorite]
Like inwoukd be curious to see a video where they run it with nothing in the plugboard, type an A. Let's say a T comes out. Ok now set the rotors bavk the original position and put in a plug from T to A... ah wait...there it is (at least part of it) ok so since the circuit goes through the plugboard first when you put in an A it would now become a T before going through the rotors so it wont come back to the T next time on that setting. It will go somewhere else.
But ok, what prevents that T from coming back as a T which then becomes an A... I mean the original problem again. Still unsolved, just that my idea for testing "what happens if..." doesn't work.
posted by If only I had a penguin... at 9:26 AM on December 27 [1 favorite]
What was the mechanism that prevented a letter from being coded as itself?
The reflector, at least as implemented in Enigma; it always leads to another letter. The British version did allow a letter to end up as itself, which made it harder to crack.
posted by ChurchHatesTucker at 9:40 AM on December 27 [1 favorite]
The reflector, at least as implemented in Enigma; it always leads to another letter. The British version did allow a letter to end up as itself, which made it harder to crack.
posted by ChurchHatesTucker at 9:40 AM on December 27 [1 favorite]
Ok...rewatching the videos with my son who is really interested, I get it now.. its because of the reflector. So if forget the plugboard (dealtvwith above)...thinking of only of the Rotors, if the first rotors connects A-> T the second connects T->Q and the third connects Q->N, let's say... so the only way you're going to get an A coming out is if on the return trip through the rotors the second rotors passes a T to the third rotor. The only way the second rotor does that is if the first rotors passes it a Q . The only way the first rotors does that is if the reflector passes it an N. So as long as the reflector does a scramble, the same letter will never come back out of the rotors.
posted by If only I had a penguin... at 9:43 AM on December 27
posted by If only I had a penguin... at 9:43 AM on December 27
« Older Zelda Day 2024 | Internet use for people aged 50 or above linked to... Newer »
Diffie-Hellman key exchange explained using paint
posted by CookTing at 8:02 PM on December 26 [3 favorites]