Re: encryption with pi

From: Tom St Denis (tomstdenis_at_iahu.ca)
Date: 07/30/04 

Date: Thu, 29 Jul 2004 23:54:27 GMT

vedaal@hush.com wrote:

> can the following simple substitution cipher using pi, be made secure
> ?
>
> C = P rot [ s(l+n) mod 256 ]
>
> where 's' is a substring of pi of length k, beginning at digit 'l',
> and 'n' is the sequenced character of the plaintext

<snip>...

What does "rot" mean anyways?

That aside... finding digits of Pi is **SLOW** and consumes a heck lot of
memory. That and who says identifying a sequence as part of "Pi" is a hard
problem? Who knows there may be a sub-exponential algorithm that given L
digits of pi you can determine where the first one was.

Alternatively who says given L digits of Pi you can't predict the next with
more than uniform probability?

Quite frankly the idea is a bit short sighted. It's inefficient and has no
provable properties other than it's not a simple pattern.

Tom