On the puzzle page of the New York Times, in addition to the daily crossword, are a 4×4 and a 6×6 KenKen puzzle. The 4×4 is simple enough that to make it interesting I do it in my head, and then write down the answers in order (the top row left to right, then the second row, etc).

It has occurred to me that this would make a diabolical secret code to use in a spy novel. Spy number one picks up the daily paper, scans the puzzle page, and then encrypts that day’s sensitive message with some substitution code that uses the sequence 1,3,2,4 or 3,2,4,1 as its encryption key, or whatever permutation of 1,2,3,4 forms the top row of that day’s KenKen.

Spy number two just needs to look at the daily paper, find the same key, and decode the message. Since the code would change every day, nobody else would be able to figure out the coding scheme just by looking at the daily messages that passed between the two intrepid spies.

There’s nothing new about the basic idea. To form a secure code, you and I just need to settle on any key that is known to both you and me, but not to anybody else. There are well-known mathematical techniques to create secure codes, such as PGP encryption, but they generally require access to a computer. There is something charming, in an old-fashioned way, about being able to create a secure message without computation.

So I started pondering, given that you and I both get the daily paper, how we might use the paper as a source for encrypting and decrypting messages, without anybody else being able to figure out our code. We could, for example, agree to use the first line of the second article in the Sports pages.

Yet in this age of search engines, just about any rule that simple could be cracked by a dedicated computer hacker with access to the daily newspaper. Could we charmingly old-fashioned spies, with our copy of the Times rolled up under the arm of our gray-tweed suit, figure out a clever way to use the daily paper as a shared key, without our code being cracked by those young whippersnappers with their infernal computing machines?

Reminds me of this tangentially related xkcd cartoon: http://xkcd.com/936/. I’m not sure exactly what lesson we can apply from that to this problem, but it seems like there should be one.

By using something quintessential human: a sense of esthetics. We could agree to use the caption of the picture with the worst dressed person. this, however, incurs the risk of having spies with different tastes.

As usual, XKCD is awesome.