One could increase the difficulty of decryption by three or four
doublings by intermixing random data with plaintext in a message.
Here's the least stupid method I can think of: the first character in
a message is the start of text (SOT) character. The second character
in a message is the end of text (EOT) character. The message itself
consists of random data intermixed with plaintext prefixed by SOT and
suffixed with EOT. An EOT outside of plaintext stands for itself. An
SOT inside plaintext stands for itself. This method can encode
arbitrary plaintext. By implication, the random data does not contain
an SOT nor EOT.
Instead of being able to look at a fixed point in the encrypted text
for plain text, it's necessary to examine the entire text. The
cryptanalyst gets a clue that they should continue if they look at
enough of the random text without finding SOT or EOT characters. They
get a clue they should stop if the first two characters are identical,
but that's only 1/256 probability.
This mainly serves to increase the complexity and expense of
decryption engines.
--
-russ nelson <[EMAIL PROTECTED]> http://russnelson.com
Crynwr sells support for free software | PGPok | No intellectual property
521 Pleasant Valley Rd. | +1 315 268 1925 voice | rights were harmed in the
Potsdam, NY 13676-3213 | +1 315 268 9201 FAX | creation of this message.
