Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Posting-Version: version B 2.10.1 6/24/83; site deepthot.UUCP Path: utzoo!watmath!deepthot!julian From: julian@deepthot.UUCP (Julian Davies) Newsgroups: ont.micro Subject: Re: flame about random numbers Message-ID: <435@deepthot.UUCP> Date: Mon, 7-Jan-85 20:00:42 EST Article-I.D.: deepthot.435 Posted: Mon Jan 7 20:00:42 1985 Date-Received: Tue, 8-Jan-85 07:44:46 EST References: <336@utai.UUCP> Organization: UWO CS, London Canada Lines: 7 I'd suggest a random number generator that *depends* on random human i/o delays to give random outputs is asking for trouble. For instance, any kind of simulation program is likely to get hurt. The generator ought to produce reasonably pseudo-random numbers even when there is no user-input variability. It isn't as if the technology isn't well understood. There's half a book by Knuth on the subject. (Semi-numerical Algorithms).