Re: [Help-gsl] Wolfram EllipticK vs gsl_sf_ellint_Kcomp

[Ruben Van Boxem]

2010/7/27 Ruben Van Boxem <address@hidden>

> A small extra that introduces quite a big problem for me: the
> gsl_fs_ellint_Kcomp function takes the sqrt of the argument of Wolfram's
> EllipticK, meaning that while Wolfram's EllipticK can perfectly handle
> things like
>
> EllipticK(-.4)
>>
>
> I see no (easy) way of doing such a thing in terms of GSL's elliptic
> functions.
>
> There may be a workaround for this, but I'm not too well versed in these
> things, the function popped up in an analytical solution to an integral I
> have to calculate.
>
> Thanks!
>
> Ruben
>

Sorry for the double post, but ALGLIB (GPL) can handle a negative
input<http://www.alglib.net/specialfunctions/ellipticintegrals.php>,
so for now, I've switched over to that library for the calculation of
EllipticK.