🚀 go-pugleaf

    Thank you to everyone who has been so helpful on sci.crypt on 
helping me learn and point out flaws in my system (all zero of you.)

    I spent the entire day checking patterns in mcrypt and I've
determined that the bit series can be summed up as:

    512
     E  (7i+1) = X    whereas   key = ((8+X)^2) * (2^16) 
     2          

    For a grand total of

    55,446,520,257,839,104 permutations compared to 
       140,737,488,355,328 permutations in DES.

    I spent the whole day writing 5000 digit floating points and
integers for my 386 to verify these facts, and explaining my theories
to countless (okay, 3) bored professors that will no doubt sleep
better tonight due to my questions.  BE MY GUEST to look at the 
methods and software I used to prove it.  Available on request. 

    Admittably, I missed my mark of 2^6044.  Still, it is a hellovalot
faster than DES to use and takes a hellovalot longer then DES to break.
I've going to experiment raising this permutation set by increasing the
salt and by by increasing the block size.  I've sure that it can still
be reasonably used at a value near 2^120.  I'm considering using a tree
to raise my 2^16 salt to a 2^32 salt, that way I don't have to risk
errors having lost hash codes.  Any suggestions on this?

    -- Eric Knight

(Does being a cryptologist mean nobody knows what your doing?)