On Fri, Jan 24, 2014 at 2:00 PM, Jim Bell <jamesdbell8@yahoo.com> wrote:
 
> Consider this:  Suppose I handed you the digits of pi, the digits from the
> millionth digit to the two-millionth digit, and I asked you to determine if they
> are 'random'.  By many tests, you'd conclude that they are random.  (Or,
> at least 'normal'  http://en.wikipedia.org/wiki/Normal_numbers  )   But, in
> reality they are highly non-random, precisely because they are a million
> sequential digits of pi.  But you wouldn't know that, if you didn't know that.

Practically, would it matter? Maybe.
 
If an attacker knew that you were using Pi as your "random" stream, I guess that would reduce your "random" stream to a stream cypher with a key of about 24 bits.
 
There are a lot of random-appearing number sequences. Are there enough to add a significant number of bits to the effective key? Against an attacker with the resources to compute and store the first billion digits of a lot of sequences?
 
Meh. I'd started this response with the plan to argue that a slice of Pi is good enough for practical purposes, but I convinced myself otherwise. It's only good enough for security-by-obscurity. Meh.


--
Neca eos omnes. Deus suos agnoscet. -- Arnaud-Amaury, 1209