Fractals, Cellular Automata, and Encryption

Mutant Rob wlkngowl at unix.asb.com
Fri Mar 8 03:27:22 PST 1996


-----BEGIN PGP SIGNED MESSAGE-----

Interesting point. I've dabbled with CA a little bit, though not in terms
of crypto.  I think at best CA can be used for stream ciphers... and
a big problem is that people assume it's a totally new form of computing,
when in fact it's only a different form, and anything that can be done
with a conventional formula-based scheme can be done with CA and visa-
versa... the pitfall is that one can get all caught up in how chaotic CA
behaves and lose sight that the same thing can be done in a formula, and
that possibly it can be easily broken.

Fractals are interesting. I've thought about using the Julia-set
iterations as a form of crypto (or for that matter, recursive methods
in general)... maybe a kind of block cipher that works with complex
numbers, but using the words as fractions rather than whole numbers.
Perhaps using 64-bit binary fractions, and iterating X = X^2 + C, where
the iteration count and C are keys... using the result as a kind of
stream cipher.  Problem is it would be slow on most machines.

I've also thought about genetic algorithms.  An interesting ideal would
be a genetic algorithm that operated on plaintext, key, ciphertext but
would be self-analyzing and evolve itself in ways to make cryptanalysis
difficult.

- --Rob
- ---
[This message has been signed by an auto-signing service.  A valid signature
means only that it has been received at the address corresponding to the
signature and forwarded.]

-----BEGIN PGP SIGNATURE-----
Version: 2.6.2
Comment: Gratis auto-signing service

iQBFAwUBMT+FpSoZzwIn1bdtAQHg9QF+OhSfJi0WIPB1Lqg5ne7f8mYYvow7yl5k
0gJh0KHaCEJZUcwhmRZ1uWlDlExcx+Q/
=sgLv
-----END PGP SIGNATURE-----






More information about the cypherpunks-legacy mailing list