home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: sci.math.stat
- Path: sparky!uunet!utcsri!torn!news.ccs.queensu.ca!mast.queensu.ca!dmurdoch
- From: dmurdoch@mast.queensu.ca (Duncan Murdoch)
- Subject: Re: Generating a set of random bits
- Message-ID: <dmurdoch.62.714343698@mast.queensu.ca>
- Lines: 30
- Sender: news@knot.ccs.queensu.ca (Netnews control)
- Organization: Queen's University
- References: <36984@sdcc12.ucsd.edu> <1992Aug19.122345.27772@cl.cam.ac.uk> <1992Aug20.195901.2259@cc.tut.fi>
- Date: Thu, 20 Aug 1992 20:48:18 GMT
-
- In article <1992Aug20.195901.2259@cc.tut.fi> vattu@cc.tut.fi (Vattulainen Ilpo) writes:
- >>nmm@cl.cam.ac.uk (Nick Maclaren) writes:
- >>Standard warnings:
- >>
- >> 2) Don't trust Numerical Recipes - about half of its techniques are
- >>reliable, but the others contain serious traps.
- >
- > I agree.
- ...
- > are. But - do not use RAN3 (in Numerical Recipes ), if you
- > want random bit-wise properties... This is due to my own
- > experiences :-/
-
- I think both people I've quoted above are being somewhat unfair to Numerical
- Recipes. Perhaps it has changed in later editions, but in the original
- edition, it plainly says that none of the routines in the book are
- applicable to this very problem, and refers the reader to Knuth. If you
- were burned by using RAN3 for a vector of random bits, you can't really blame
- Press et al.
-
- More generally, I think it's true of advice from most books (and certainly
- of that received over the net) that at least half of it contains traps for
- the unwary. Saying this about Numerical Recipes in particular doesn't
- really convey any information.
-
- Thanks to both of the above for posting useful answers, as well as what
- I've picked out above.
-
- Duncan Murdoch
- dmurdoch@mast.queensu.ca
-