[Cryptography] Holiday-themed decode contest

[Brad Klee]

> (Somewhat off-topic, apart from the fact that the density of information
> would make it a great subliminal channel).

The relevance is that we can always calculate ratios of subspace sizes,
including signal size : block size. In the contest, it changes per
dimension,
(up to a factor of two) 3D is 1:64, 2D 1:16 and 1D 1:1 (without symmetry).

However, signal size may need to be reduced if we are assuming 16^n
blocks encode to a set of n characters (not necessarily true). Interesting
combinatorics to compare size of n-symbol codebooks if the alphabet
is CJK or western a-Z...

Here's an academics joke since you were nice enough to respond:

If you're definitions *require* 1:1 signal size : block size, obviously 3D
case
*can not* be properly called a Feistel Cipher. 1D case passes this test, but
why so much symmetry during the mixing phase? The answer is:

(96B1~*隱*) All of the above!

If no one gets a decode, I guess we could discuss whether or not definitions
should ever be export controlled? In this case, I'm sure they have better
options at asiacrypt, but in general I think "No". Should they export
control
PhD dissertations? Same answer, "No"! Open Access Only!

It's funny to me to imagine a voxelized cryptogram floating ominously
through
the Sandbox or in the ocean around a townscraper model, but the 3D algorithm
probably wouldn't be very useful in practice. I haven't proven weaknesses of
the 1D case, but it computes relatively fast and the unicode keymap is easy.

> I've never figured out how they get some of the stuff on there, I mean
why
> is today "an auspicious day to dig a ditch"?

My guess is that Xinjiang isn't that different than Cool Hand Luke. "If
your
brother's frozen fecal matter is in the people's latrine, your time could
be
up, and have to dig, man."

--Brad
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://www.metzdowd.com/pipermail/cryptography/attachments/20220104/8aa09be8/attachment.htm>