[SciPy-Dev] Use NPY_INFINITY in cephes code in scipy.special?
Sat Dec 4 14:48:25 CST 2010
On Sat, 04 Dec 2010 14:40:13 -0600, Warren Weckesser wrote:
> scipy.special.zeta(1.0, 1) returns 1.7976931348623157e+308 to indicate
> infinity. Here's the relevant snippet from zeta.c:
Yes, proper infinities should be used instead of MAXNUM. I don't think
there's any valid mathematical reason for returning a big but finite
result at points where the function is numerically infinite. Fixing these
systematically could be useful.
More information about the SciPy-Dev