Merge Marsaglia-Tsang routines with GSL?

John D Lamb J.D.Lamb@btinternet.com
Fri Jan 6 19:24:00 GMT 2006


Would it be useful to merge these routines (currently extensions)
directly into GSL?

I'm happy to help with the coding, document what the code does in LaTeX
and write some suitable code for a test program.

I would suggest some names like gsl_ran_gaussian_mt and gsl_ran_gamma_mt
to distinguish the code from the existing functions. They could even
replace the existing functions, though there are two differences: the
sequence of generated variates is different and so could conceivably
change the results of old code; the Marsaglia-Tsang generators assume a
32-bit random number generator.

I assume Jochen Voss would be happy to contribute his code. I'm
certainly happy to contribute mine and transfer copyrights,
attributions, etc.

-- 
JDL



More information about the Gsl-discuss mailing list