Date: 02 Nov 2001 10:15:06 -0800 From: swear@blarg.net (Gary W. Swearingen) To: freebsd-chat@FreeBSD.ORG Subject: Re: Unix Philosophers Please! Message-ID: <mlofmlca51.fml@localhost.localdomain> In-Reply-To: <20011101214720.G4360@blossom.cjclark.org> References: <3BE08283.EC81A8ED@math.missouri.edu> <20011031170629.C865-100000@coredump.scriptkiddie.org> <20011101214720.G4360@blossom.cjclark.org>
next in thread | previous in thread | raw e-mail | index | archive | help
"Crist J. Clark" <cristjc@earthlink.net> writes: [ about pi digits as random numbers ] > > If you were to run all of your statistical tests on that set of > numbers, it would appear to be random. But it is not. If you tell me > any arbitrary position in the bit stream, I can tell you what the next > value will be. Not random. Just a side-note about the opposite "gotch" to also watch out for: "0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0" could be the output of a truly random number generator and it could thus be said to be a random number sequence. You can not tell me what the next value wil be. Random. (You will never get that sequence from a common psuedo-random number generator.) So when some software ask you for a random number sequence, you should, in most cases, not only feed it random numbers, but numbers that are designed to produce numbers that look random even though they are not or those that have passed statistical tests for randomness. So it will often be better to use numbers from random(3) (maybe seeded by numbers from /dev/random) than those from /dev/random. (Unless those devices' output has been designed to also pass statistical tests in all cases, rather than seeming to be truly random (which won't pass statistical tests in all cases). Anyone know?) To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-chat" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?mlofmlca51.fml>