06 January, 2006

Marian Rejewski on Wikipedia

The Wikipedia article on Polish mathematician Marian Rejewski is currently going through the process to become a so-called "Featured Article", vetted entries that are eventually promoted on the main page and elsewhere.

Feel free to read the article and add any comments. The page for the Featuring process can be found here.

Despite having something of a girl's name, Rejewski was a veritable geek hero, solving the wiring of the German Enigma machine using some funky maths. It's more than possible that, without Rejewski's results, the British at Bletchley Park would have had little, if any, success with Enigma.

Rejewski also had a more exciting wartime experience than his counterparts at Bletchley Park, who were all off playing rounders in relative safety. From September 1940, Rejewski worked with a small unit on breaking Nazi ciphers from within occupied (well, Vichy) France, under the continual threat of discovery and arrest. After being nearly discovered by a detector van equipped with a radio antenna, the unit was evacuated. Aided by the French resistance, Rejewski worked his way to the border and attempted to cross over into Spain. He didn't have much luck at this point: robbed by his guide at gunpoint, captured by Spanish police only hours after crossing the border, and then interred for three months in prison.

After his release, he made his way to Britain. You might assume he would join the codebreakers at Bletchley Park on Enigma, since their work had built on his. Instead, he was assigned to a unit working on low-level codes; a decision that one former British codebreaker described as "like using racehorses to pull wagons". It's difficult to tell what motivated this apparent injustice -- perhaps a need for security, particularly since the British had developed more advanced techniques than those of Rejewski, and the future of Poland itself was uncertain. It's also possible that the relevant authorities did not know of the Polish contribution to the British work. Most of the few that had known were no longer immediately involved with Bletchley Park.

Regardless, Rejewski deserves to be remembered as one of the "greats" in crypto history.

05 January, 2006

National Cryptologic Museum

Adjacent to the NSA headquarters in Maryland, USA is the National Cryptologic Museum, a museum of cryptology and NSA history open to the public. It's surely second only to Bletchley Park as the ultimate day out for crypto geeks (if geeks had days out, that is).

Regrettably, I've not had chance to visit -- it's a bit of a trek from the UK -- but the next best thing is a superb set of photos posted to Flickr by Austin Mills which document a large proportion of the museum's exhibits. The photos are high enough resolution to be able to read the captions and see the details of the various machines.

01 January, 2006

From the archives: Mercury

I had chance to spend a day at the UK National Archives a couple of weeks ago, and came across a file (AVIA 65/977) discussing cash awards to the designers of Mercury, an on-line cipher machine used by the RAF from 1950 to the 1960s. I wrote up all the information I could glean about this machine in a Wikipedia article.

Mercury embodies a similar principle to the US SIGABA (ECM Mk II) machine: some rotors control the movement of other rotors. It would appear that Mercury was an independent reinvention of the concept, as the ECM was around a decade older. The ECM concept had been kept secret from the UK by the US, and a Combined Cipher Machine (CCM) was used instead for inter-allied communication.

Another feature of Mercury appears to be that of "double rotors". Exactly how this worked wasn't clear from the PRO documents, but one plausible guess is that there were two independent wirings inside each rotor core, and that these wirings could be set in 26 offsets from each other, and optionally reversed. The current would pass once through the "inner" wiring, and then back again through the "outer" wiring. Ordinary Typex rotors had a "double" contact feature used to improve the reliability of the electrical contacts; this could have been pressed into service as a double rotor feature. Each pair of double rotors would, of course, step together.