<[EMAIL PROTECTED]> wrote: > On 24 Apr 2002 at 17:41, David Howe wrote: > > its probably a better (if much slower) stream cypher than most currently in > > use; I can't think of any that have larger than a 256 internal state, and > > that implies a 2^256 step cycle at best; for pi to be worse, it would have > > to have less than 2^256 digits. > This is putting sillines on top of silliness. It's true that in principle > that the decimal expansion of pi has an infinite number of digits, > but any practical implementation of a PRNG based on pi > would still have to have a finite number of accessable states. Indeed my point (the mentioned hardware implimentation limitations) - however, you don't need an infinite pi - a prng based on a subset that has 2^257 bits of the sequence has by definition a longer cycle time than a 256 state prng.
> Conversely, a PRNG whose cycle is "only" 2^256 bits long > will never repeat itself during the lifetime of the device, or > the lifetime of the universe for that matter. which is why a subset is sufficient.