I've been working through the ramifications of my new algorithm and it

seems that it can be used as an encryption busting tool.

Take the fictious case of a public encryption key which is composite

with two prime divisors, say 245969 (scaled small for the sake of

example).

Feeding 245969 into my algorithm gives 1073 which further factors down

to 29 * 27. It follows from this that one factor also divides 2^29-1

and one divides 2^37 * 1. This of course narrows down the search into

smaller chunks, hence decreasing the time to cracking the encryption.

The factors in this example are actually 223 and 1103.

Furthermore, given one factor of a Mersennem it is now possible to

derive both the Mersenne itself and the second factor.

Hopefully I won't get the NSA knocking at my door because of this.

Kevin.