06-12-2003, 04:34 AM
Quote:Mango: I have been thinking of using CA (2D, though) to encrypt files. Using a key to generate another, larger key and then adding it into the plaintext is possibly unbeatable.
What do you mean by "The key is updated as it is used"??????????
Agamemnus...I'm assuming "The key is updated as it is used" came from the readme file? Basically, my key is a subset of the final state of the "rule 30" 1-D CA, seeded with the password. In V 3.11, I constructed it so that the key would be variable length, but in the 500-1000 bit range. Therefore, if a file longer than 1000 bits is being encrypted, the key must either repeate (generally a no no) or be updated at each iteration. As I wrote 3.11, the key is updated using a modification of the rule 30 automata used to generate the key from the password. This turns out to be a problem, *IF* an attacker has partial knowledge about the contents of an encrypted file...it is of no help if the attacker is in the dark, but helps him if he is already partially in...I've changed my key update method in the current iteration.
RE: 1d vs 2d CA....read Wolfram's book...it really doesn't buy you anything to use a 2d...once the threshhold of complexity that will produce "intractiblly random-appearing behavior" is reached, it doesn't buy you *anything* to try to make it *more random appearing*. In fact, you actually run the risk of deluding yourself...if you use a CA of higher dimmensionallity you may end up with a repeating pattern that you don't reccognize...Make sure whatever 2d CA you use is non-reversable, and produces non-nesting patterns...
I've actually studied and learned a BUNCH in the last 6-mos on this topic...if you want to discuss what you are doing, or details of my scheme, e-mail me and I'll gladly respond...
did you try my program???
did it work?
I hope so...it gives great pleasure to find out that someone else was actually able to successfully run your code... ;-)
It should produce random-looking data irregardless of the regularity of the input data-stream or the password chosen...In fact, I submit that the output of *any* file encoded using jcrypt 3.1 with *any* password is a better PRNG (pseudo random number generator) than the RND function in QB 4.5!!!
Cheers...I hope I answered your Q regarding key update...