This is the mail archive of the
gsl-discuss@sources.redhat.com
mailing list for the GSL project.
Re: Landau distribution
Hi Brian,
In Shorr's original paper on the Landau distribution he does quote the
numerical precision I mentioned before. I'm aware that work has been
done on this since then - in particular, I know about the CERN library
(used it every day and late into the night as a grad student). The CERN
routine (called RANLAN) uses, "Approximation by rational functions. For
reason of speed, RANLAN proceeds mainly by table look-up and quadratic
interpolation."
What I didn't know was that the CERN library has been GPL'ed. It wasn't
always. Here's that info:
http://wwwinfo.cern.ch/asd/cernlib/conditions.html
Therefore, I'd suggest recoding the original Fortran source into C as
the way to go. I can look into this. That was the tack taken by the
developers of ROOT (Rene Brun, Fons Rademakers) in their C++
implementation:
http://root.cern.ch/root/html/src/TRandom.cxx.html#TRandom:Landau
There are lots of other things in the CERN code that'd be worth mining
for use in GSL.
Dave
--
David Morrison Brookhaven National Laboratory phone: 631-344-5840
Physics Department, Bldg 510 C fax: 631-344-3253
Upton, NY 11973-5000 email: dave@bnl.gov