From: gtoal@an-teallach.com (Graham Toal)
Date: Wed, 27 Jul 94 13:07:45 PDT
To: cypherpunks@toad.com
Subject: Re: Cryptosplit note
Message-ID: <199407272006.VAA07358@an-teallach.com>
MIME-Version: 1.0
Content-Type: text/plain


	   It uses rand() when it needs random numbers for the
	coefficients of the polynomial. I don't know what kind of
	security risk that poses, but it really should be using something
	better.

It definitely should.  It will be trivially crackable. (Well, trivally
after someone writes the program I mean, which is more tedious than
difficult).

(Next question to determine *how* tedious... what do you do to initialise
the random sequence? - oh, never mind, I should look at the code, except
it's probably not worth the effort now.  Just make sure people only
use it for fun and not serious data hiding.  Ask on sci.crypt if anyone
wants to take a crack at decrypting the exored data, given the source
code.)

	  Where can I get Blum-Blum-Shub source or documentation on the
	algorithm?

Can't help on that one, sorry.  They don't tell us nasty furriners.

g