>Harvey, this looks very good, but what properties of the set of primes >are actually needed for this result? Will any infinite set of a similar >asymptotic density work? I use only that no product of primes is equaled to any other product of primes.