Article View: comp.sys.amiga.tech
Article #1473Re: Random #s (was Re: ARexx pseudo-random number generator)
From: Tom Bruhns
Date: Sat, 30 Sep 1989 00:26
Date: Sat, 30 Sep 1989 00:26
27 lines
1545 bytes
1545 bytes
andrew@teslab.lab.OZ (Andrew Phillips 289 8712) writes:
>.
>.
>.
>Another interesting idea that I saw in sci.crypt recently is to test a PRNG
>by drawing pixels on a screen. A good one will make something that looks
>like snow on a TV. The eye can very easily detect some patterns which are
>caused by a poor PRNG. ...
A way to make this extremely sensitive for at least some schemes:
Method A: Plot two sequential points as x and y on the screen. But
only plot them if the first of the two is within a range which is very
narrow compared with the range of the generated numbers. Obviously,
spread its axis so that narrow range fills the screen.
Method B: Draw numbers till one falls in a narrow range, as in method
A. Draw two more, and plot them as X-Y.
Method A will display problems in congruential generators; method B is
nice to see how you have done on a binary shift register. Expect to see
a pattern like a sawtooth wave for A and a pair of interleaved triangle
waves for B, on the suggested generator types...
>--
>Andrew Phillips (andrew@teslab.lab.oz{.au}) Ph. +61 (Aust) 2 (Sydney) 289 8712
>----------
Technical Details
Message-ID:
<170008@hplsla.HP.COM>Path:
arc1-ca.pugleaf.net!utzoo!attcan!uunet!cs.utexas.edu!tut.cis.ohio-state.edu!ucbvax!hplabs!hp-pcd!hplsla!tombReferences:
<194@teslab.lab.OZ>Show Full Headers
Complete Article Headers
Path: arc1-ca.pugleaf.net!utzoo!attcan!uunet!cs.utexas.edu!tut.cis.ohio-state.edu!ucbvax!hplabs!hp-pcd!hplsla!tomb From: tomb@hplsla.HP.COM (Tom Bruhns) Newsgroups: comp.sys.amiga.tech Subject: Re: Random #s (was Re: ARexx pseudo-random number generator) Message-ID: <170008@hplsla.HP.COM> Date: 30 Sep 89 00:26:08 GMT References: <194@teslab.lab.OZ> Organization: HP Lake Stevens, WA Lines: 27