In article <0.ef56b5652decd19bb478.20121128013501GMT.firstname.lastname@example.org>, Ben Bacarisse <email@example.com> wrote:
> Good quality, hardware-generated random number sequences (if our current > understanding of quantum effects is correct) are random in a different > way to the digits of pi. It helps if the terminology is be able to > distinguish between them.
I do not see how quantum effects can be used to generate random sequences. Coherent systems are stable and highly, if not perfectly, predictable.
Hardware generated random sequences usually read Schottky noise off some device (a sound card in a computer) and use that. This can be modeled using entirely classical physics.
Given the sampling frequency and bin size and nuber of bins and the physics of Schottky noise and the particular device, a lower bound on the information stream of the readings can be made. Say, pessimistically, the reading can be one of exactly two states: A and B with Pr(A) = 0.99, Pr(B) = 0.01. The information is 0.08 bits per reading. If we take 2000 readings we have 160 bits of information. Hash the 2000 readings into a 160 bit hash. Presto! Chango! 160 random bits.