Hacker News new | past | comments | ask | show | jobs | submit login

My understanding is that the Polish team cracked the three rotor Enigma. Turing & co extended that work to four rotor plus plug board.



They reverse-engineered the machine sight-unseen (determining the permutation wired into each rotor) and cracked the original way that per-message settings were sent.

A daily key was used to encrypt the "indicator" which determined the initial state of the machine for the rest of the message. Initially, the indicator was repeated (ABCABC); it turns out that if you collect a bunch of indicators, this repeated structure allows you deduce the rotor wiring and also to determine what the key of the day is and thus read all the traffic.

https://en.wikipedia.org/wiki/Cryptanalysis_of_the_Enigma#Re...

Gradual changes to German practices (different key management, additional rotors to choose from, etc.) increased the work required to the point where it was beyond the limited means of the Polish cryptographers alone, but shortly before the start of WWII they started working with French and British cryptographers and gave Bletchley Park a running start.


Exactly. The Polish cryptanalysts cracked the cypher, and even built the prototype bombe. But the technology had to be scaled up for production.

That scaleup is the real story of WWII cryptanalysis. Prior to WWII, cryptanalysis had been a few people in offices working by hand. During WWII, on the Allied side it became an industrial operation. Bombe key-testers were built in quantity by the British Tabulating Machine Company and National Cash Register. They were spread around England, so that no one bombing raid could take out more than a few. The US had a big machine cryptanalysis operation at Arlington Hall, and another one in Hawaii. There were radio intercept stations all over Britain, and the US had ones on Pacific islands.

See NSA's history of WWII cryptanalysis.[1] They credit the Poles; Turing gets a mention. Gordon Welchman's contribution, the "diagonal board", which reduced the effort required by a factor of 26, is seen as more significant.

Turing didn't like the NCR machines. The British bombes just stopped when they got a hit, and someone had to copy down all the wheel positions by hand. NCR's machine was faster and had a printer. When the machine got a hit, it would go past the hit (because the contact drums were spinning too fast to stop instantly), back up, slowly go foward to find the hit again, stop, print the wheel positions, and restart.

More automation followed. The Enigma was a hand machine, and someone had to copy the output by hand as the lamps lit up. The US cryptanalysis operation had a machine driven from paper tape to do that job, once the keys had been cracked.

Turing thought this automation was unnecessary.[2] But once it was all working, info about German submarine positions was flowing through the system rapidly, and within months, so many German submarines had been sunk that Donetz abandoned the Battle of the Atlantic.

It took a huge industrial operation to bring this off.

[1] https://www.nsa.gov/about/cryptologic_heritage/center_crypt_... [2] https://www.nsa.gov/about/_files/cryptologic_heritage/public...


According to the wikipedia article[1], the Polish lacked resources to automate decryption once the Germans upgraded Enigma.

[1] https://en.wikipedia.org/wiki/Cryptanalysis_of_the_Enigma#Po...


That's my understanding too. They got the basics, Turing worked on a more sophisticated problem derived from that.




Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: