On Wed, 30 Jun 1999, H. D. Vinod wrote:
> Does the new version of S-PLUS solve some of the seriuos
> numerical accuracy problems pointed out in McCullough (1999)?
Which new version did you have in mind: S-PLUS 5.1 or 2000? The RNG is
unchanged in both, and quite rightly so.
> For an academic user like me, too many new versions are not good.
> We wish they will simply fix the older versions.
Who is the `we' that you are speaking for here? Isn't it a bit arrogant to
assume to speak for all academic users? Some of use in world-class
universities try to keep up with the latest statistical methods, even to
teach them to our students.
> For those who may have missed The American Statistician article , you may be
> interested to know that B. D. McCullough has assessed the numerical
> reliability of S-PLUS, SAS, and SPSS. The articles can be downloaded from
> the TAS website http://www.amstat.org/publications/tas/
>
> It seems that the random number generator(RNG) for s-plus fails four of the
> DIEHARD tests and
Yes, is that relevant? All PRNGs (they are _not_ RNGs) fail some tests, and
some of the other packages are just using PRNGs which are tuned to pass
DIEHARD. You can prove almost anything that way: the S-PLUS PRNG certainly
does not have `seriuos numerical accuracy problems': Bruce McCullough's
view is not a definitive one, but just one author's opinion.
--
Brian D. Ripley, ripley@stats.ox.ac.uk
Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/
University of Oxford, Tel: +44 1865 272861 (self)
1 South Parks Road, +44 1865 272860 (secr)
Oxford OX1 3TG, UK Fax: +44 1865 272595
-----------------------------------------------------------------------
This message was distributed by s-news@wubios.wustl.edu. To unsubscribe
send e-mail to s-news-request@wubios.wustl.edu with the BODY of the
message: unsubscribe s-news
|