I have no view of the usefulness of physical RNGs for game playing, dance-move generation, and the like.
They are, however, all but useless in programs, where they have a long, highly problematic usage history. IBM's first models of the circa-1958 SAGE version of MIT's Whirlwind computer were equipped with shot-noise-based physical RNGs, but they were discarded early on. The sequence of values generated by a physical RNG is not reproducible, and programs that use one are for this reason all but impossible to test and debug. Note also that the acronym PRNG, in wide use in the gaming community, is better avoided. Its initial P is ambiguous, can be interpreted as standing for either 'physical' or 'pseudo-'; one of my very young students even guessed that it stood for 'programmed'. John Gilmore, Ashland, MA 01721 - USA ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
