update to sphere.c in randist

[James Theiler <jt@nis.lanl.gov>]

hello GSL'ers,
  i don't really think this message merits the wide distribution of
gsl-discuss, but this is my first time using the sourceware interface
to gsl's cvs site and mailing list, so you can think of this as a
longwinded "test" message.
  the checkout, autogen, build, and cvs all worked great by the way!
i didn't build everything (which is a further tribute to the modularity
of this package), but what i did build went without a hitch.
  the only technical content of this message is that i updated the
algorithms used in randist/sphere.c; i'd been doing something like that
for another project and happened to have Knuth (the book, not the guy) on
my desk, and so used this as an excuse to try out the new gsl/sourceware
interface.  i'm not sure i'll always be this disciplined, but when adding
a new routine to gsl, i try to be generous with comments, but mostly i try
to have a reference that i can cite, so the algorithm is less of a black
box.  my vision of gsl is not so much a comprehensive math library (though
it is rapidly heading in that direction, to the credit of its developers)
as a place to go to get some good numerical source code, unencumbered by
restrictive copyright, which can then be modified to the particular
project you are working on.
  i'm sure by now mark galassi is the only one who'll have read this far,
so one last note: the old cvs repository for gsl at LANL is still active
(or at least the /n/projects/cvs/gsl directory is still there).  now that
the transition to sourceware.cygnus seems to be complete, is there any
reason to keep the old one around? 

regards to all,
jt

---------------------------------------------
James Theiler                     jt@lanl.gov
MS-D436, NIS-2, LANL        tel: 505/665-5682
Los Alamos, NM 87545        fax: 505/665-4414
----- Space and Remote Sensing Sciences -----