jon@balder.us.dell.com (Jon Boede) said:
Can anyone refute the claim that you can only generate truly random numbers using hardware? I recall seeing a proof that totally random numbers could only be generated by an infinitely large program.
Assuming a bit of leeway in interpretation, this is trivially true mathematically. There's a great von Neumann quote that goes something like "anyone who uses finite state machines to generate supposedly random numbers is, of course, living in a state of sin." Use of hardware random number generation does not automatically confer a state of grace, however. Such processes sample through an aperture and are subject to the Nyquist limit, the General Uncertainty Principle, and frequently the Central Limit Theorem as well, which is to say that you still have to mind your p's and q's quite carefully. Doug