At 06:23 AM 6/30/96 -0700, Timothy C. May wrote:

While radioactive decay is unpredictable (so are a lot of things, by the way), there are all kinds of biases that reduce the apparent entropy. Detector "dead time" is a classic one (basically, the detector can't detect counts during a post-pulse recovery time...probably not a problem at low count rates, but an example of how subtle things can sneak in).

If he has more than eight bits of timing resolution, such biases will have no affect. He is using his non uniformly distributed random number to select a uniformly distributed pseudo random number. Provided that the does not attempt to get more entropy out than he puts in, the result should be a uniformly distributed truly random number. --------------------------------------------------------------------- | We have the right to defend ourselves | http://www.jim.com/jamesd/ and our property, because of the kind | of animals that we are. True law | James A. Donald derives from this right, not from the | arbitrary power of the state. | jamesd@echeque.com