Cryptonomicon

Waterhouse spends much of the next week commuting to London for meetings at the Broadway Buildings. Whenever civilian authorities are going to be present at a meeting--especially civilians with expensive sounding accents--Colonel Chattan always shows up, and before the meeting starts, always finds some frightfully cheerful and oblique way to tell Waterhouse to keep his trap shut unless someone asks a math question. Waterhouse is not offended. He prefers it, actually, because it leaves his mind free to work on important things. During their last meeting at the Broadway Buildings, Waterhouse proved a theorem.

It takes Waterhouse about three days to figure that the meetings themselves make no sense--he reckons that there is no imaginable goal that could be furthered by what they are discussing. He even makes a few stabs at proving that this is so, using formal logic, but he is weak in this area and doesn't know enough of the underlying axioms to reach a Q.E.D.

By the end of the week, though, he has figured out that these meetings are just one ramification of the Yamamoto assassination. Winston Spencer Churchill is very fond indeed of Bletchley Park and all its works, and he places the highest priority on preserving its secrecy, but the interception of Yamamoto's airplane has blown a gaping hole in the screen of deception. The Americans responsible for this appalling gaffe are now trying to cover their asses by spreading a story that native islander spies caught wind of Yamamoto's trip and radioed the news to Guadalcanal, whence the fatal P-38s were dispatched. But the P-38s were operating at the extreme limit of their fuel range and would have had to be sent out at precisely the correct time in order to make it back to Guadalcanal, so the Japanese would have to have their heads several feet up their asses to fall for that. Winston Churchill is pissed off in the extreme, and these meetings represent a prolonged bureaucratic hissy fit intended to produce some meaningful and enduring policy shift.

Every evening after the meetings, Waterhouse takes the tube to Euston and the train to Bletchley, and sits up late working on Rudy's numbers. Alan has been working on them during the daytime, so the two of them, combining their efforts, can almost pound away on it round the clock.

Not all of the riddles are mathematical. For example, why the hell do the Germans have Rudy copying out big long numbers by hand? If the letters do indeed represent big numbers that would indicate that Dr. Rudolf von Hacklheber had been assigned to a job as a mere cipher clerk. This would not be the stupidest move ever made by a bureaucracy, but it seems unlikely. And what little intelligence they've been able to gather from Germany suggests that Rudy has in fact been given a rather important job--important enough to keep extremely secret.

Alan's hypothesis is that Waterhouse has been making an understandable but totally wrong assumption. The numbers are notciphertext. They are, rather, one-time pads that the skipper of U-553 was supposed to have used to encrypt certain messages too sensitive to go out over the regular Enigma channel. These one-time pads were, for some reason, drawn up personally by Rudy himself.

Usually, making one-time pads is just as lowly a job as enciphering messages--a job for clerks, who use decks of cards or bingo machines to choose letters at random. But Alan and Waterhouse are now operating on the assumption that this encryption scheme is a radical new invention--presumably, an invention of Rudy's--in which the pads are generated not at random but by using some mathematical algorithm.

In other words, there is some calculation, some equation that Rudy has dreamed up. You give it a value--probably the date, and possibly some other information as well, such as an arbitrary key phrase or number. You crank through the steps of the calculation, and the result is a number, some nine hundred digits long, which is three thousand binary digits, which gives you six hundred letters (enough to cover one sheet of paper) when you convert it using the Baudot code. The nine-hundred-digit decimal number, the three-thousand-digit binary number, and the six hundred letters are all the same abstract, pure number, encoded differently.

Meanwhile, your counterpart, probably on the other side of the world, is going through the same calculation and coming up with the same one time pad. When you send him a message encrypted using the day's pad, he can decipher it.

If Turing and Waterhouse can figure out how the calculation works, they can read all of these messages too.

Chapter 41 PHREAKING