Re: needed: reviewers for an implementaion of AES

ddyer_at_real-me.net
Date: 10/17/05 

Date: 17 Oct 2005 10:50:06 -0700

I did anticipate that using md5x2 in too simple a way would compromise
the key length. I'm using.
        Md5HexString(passphrase,0,digits);
        UnHexify(digits,32,dig1,16); // store the actual hash
        Md5Digest(digits,0,dig2); // hash the hexified md5
        for(i=0;i<16;i++) { *key++ = dig1[i]; *key++=dig2[i]; };
..assuming that the ascii/hex in the second hash would
add unique noise.