Re: Paul Brown on Solitiare randomness flaw?

[Thomas W. Kellar]

I wrote a rc4 based encryption/decryption program for my TI86.  You can
enter an alphabetic key (up to 256 characters). Enter some uppercase only
alphabetic plaintext, then it displays decimal numbers representing the
encryption of the plaintext.  In decrypt mode you enter your key, then the
decimal numbers previously produced and it displays your plaintext again.  
I used B. Schneier's description in _Applied Cryptography_ 2nd ed as the
basis for the rc4.  If anyone wants a copy of it, send me a note.  As a
USA citizen residing in the USA, I believe you are supposed to be
similarly situated before I can send it to you. Considering that the
key can only have 28 possible values for each octet, there is probably
not much security in it. (A..Z plus space and = sign)

Thomas Kellar