in reply to Re: Breaking double-xor encryption
in thread Breaking double-xor encryption

Yeah. Any time you tell someone they shouldn't do something (like invent their own cryptosystem), they'll try to do it anyway. Problem with crypto is that they can't tell when they've failed. Unfortunately, RC4 has its problems, too, but it's still better than rolling your own.

I have to come clean about something - I said I was attacking messages with 3*keylen ciphertext bytes, which is true but misleading. I don't have a really good general method for solving the upper bits of the keys from such a short message. I was doing that with about 7*keylen bytes, until I started focusing on specific weaknesses in CipherText for that step. I'm not telling everything here, because it only has a narrow applicablity, and I don't want Steeeeeve to plug up the problems with his cipher and come back with a new version.

And because that part of my code is ugly. Ick.

Replies are listed 'Best First'.
Re: Re: Re: Breaking double-xor encryption
by John M. Dlugosz (Monsignor) on Aug 24, 2001 at 23:50 UTC
    Yow.

    I don't even understand what he was talking about, re domain stuff!

    Sounds like he's back in the 15th century, with his point about "cat" producing different cyphertext each time it's (this one uses the appostrophe) encountered.

    —John