Newsgroups: sci.math.num-analysis >Subject: --text follows this line-- Deos anybody know which algorithm Mathematica[TM] uses to evaluate elliptic integrals of the third kind with _complex_ arguments? It seems to be OK for some special cases, but I'd like a second opinion. [And I hate black boxes ...]
In ``the usual places'' (netlib, CERNLIB, ...), I found first and second kind for complex arguments. But or the third kind only implementations for real arguments seem to exist. The published mathematical literature seems to ignore the complex case, which appears naturally after partial fractioning
Or am I just missing a transformation that would allow me to avoid the complex poles?