unforum

[mapmaker]

[PUZZLES] Correspondences

[Abraxas]

From the Archimedes thread we know that the mysterious correspondent is writing to Claudia.

When I try that on the second letter I get A*I and some instances of *A.
This leads me to believe he writes in French. But why would she write in English and he writes in French? (see the letter from the Archimedes thread where she mentions a "friend" he was writing about..."ami"? Is this a simple substitution or not?

Edit: MMM9 doesn't seem to speak for something as simple as that...

[mapmaker]

So I thought I'd bump this, give some excitement to the VaporLofts boards, and try to get people interested in this puzzle. I think it is solvable, but we need to figure out the cipher!

We know it has both numbers and letters. 36 characters. Furthermore, there are numerous one-character "words" - 0, 8, U, K, C, etc., which would imply that it's not just substitution. So what could it be?

Well, there are a number of classical ciphers that use a 5 x 5 grid of letters to encode plaintext - many of these can be adapted to use a 6 x 6 grid instead (see page 2), including Bifid, Playfair, and others. That's my best guess.

But there's a problem. Most of these are digraphic - that is, they take pairs of letters when encoding. Why is this a problem? Well, each letter begins with the same word: 04GN35H. The next letter in each case is different - M, U, and H in each letter.

Why is this an issue? Well, let's assume that each letter starts off with the same plaintext - a reasonable assumption in my mind. Let's say it is CLAUDIA, as Abraxas has smartly noted (although the problem remains even if it isn't CLAUDIA). It would start with 04=CL, GN=AU, 35=DI, which is fine. But then we get A?=H?. The odds of an A pairing three digrams with H's is very low.

Then again, I haven't tried hard to figure out how possible it is. It may be that this is a huge weakness in the way the correspondences are encrypted. But I'm doubtful a digraphic technique is going to work.

I've personally tried to do Ragbaby decryption on one of the letters, with poor results.

Anyway, I thought that this might stimulate discussion about this particular puzzle. Does anyone have any thoughts?