On Thu, Apr 14, 2005 at 10:59:48AM -0700, Ben Pfaff wrote:
Jason Stover <[EMAIL PROTECTED]> writes:
> The bisection method used in betadistinv.c to get an initial
> approximation wasn't getting close enough to the right
> answer for p=.99, a=2.5 and b=5. I fixed this, and tidied
> up a bit to make the algorithm a little more robust.
Thanks! I committed your fix, and updated the test cases.
This takes care of the problems cited by John, as far as I can
tell.
Thanks Jason.
I'm happier with that now. Note that
test/expressions/randist/{f,beta}.out are now identical to the
original versions.
J'
--
PGP Public key ID: 1024D/2DE827B3
fingerprint = 8797 A26D 0854 2EAB 0285 A290 8A67 719C 2DE8 27B3
See http://pgp.mit.edu or any PGP keyserver for public key.
pgp1bEGCzb2dE.pgp
Description: PGP signature
_______________________________________________ pspp-dev mailing list [email protected] http://lists.gnu.org/mailman/listinfo/pspp-dev
