On 9 Oct 2013, at 20:14, Erik Schnetter <schnetter@cct.lsu.edu> wrote:

We should remove all other phi integration methods.

Since this makes old results unreproducible, I don't like this idea.  How about adding parameters to separate the theta and phi directions?

Multipole::phi_integration_method = "spectral"
Multipole::theta_integration_method = ...