home *** CD-ROM | disk | FTP | other *** search
- Path: ccnet.com!not-for-mail
- From: felinoid@ccnet.com (Jeffery C McLean)
- Newsgroups: comp.lang.c
- Subject: Re: pseudo-random numbers
- Date: 19 Jan 1996 19:36:20 -0800
- Organization: CCnet Communications (510-988-7140 guest)
- Message-ID: <4dpnvk$9h8@ccnet3.ccnet.com>
- References: <17709D420S86.JJSTEP00@ukcc.uky.edu> <4d3ijv$9fl@news.iag.net> <DLE110.Ixp@thinkage.on.ca> <1771210427S86.JJSTEP00@ukcc.uky.edu> <4do5e7$9qr@ccnet3.ccnet.com> <Pine.OSF.3.91.960119114411.18779D-100000@io.UWinnipeg.ca>
- NNTP-Posting-Host: ccnet3
-
- Bill Simpson (wsimpson@uwinnipeg.ca) wrote:
- : > But I alwase felt that using a shorted or flaky transister.. one that
- : > produces random results.... would make a quality random number generator
- : George marsaglia, random number expert, says that all physical devices
- : he has tested FAIL his tests for randomness. That is they are worse than
- : deterministic computer algorithms.
-
- You mean someone will have to acualy design a falty curcut
- for the precise function of being random?
- (please note the place I get this idea from is a random number genorater
- in the SID chip on the C64.. a single bit a read to get a random number.
- It has been very relyable for me and so I never learned to produce
- psudo random numbers untill I went to Dos... and I never even bothered
- with making randome numbers in CP/M)
- My random number macanisum was not very good... It used
- both an advanced mathmatic encryption of the time and date plus
- a twist of timing the vido sync pulse... It didn't work out to well.
- If someone were to make a flaky curcut on purpouse for the
- fucntion of random numbers I would segest designing it as an anolog
- as more varables are posable that way... then the result can
- be converted to digital for reading.
- Jeff
- --
- felinoid@ccnet.com
- http://www.ccnet.com/~felinoid/
-