Sampling sphere with 3 parameters? about posym

[kovalp]

I formulate my ideas as issues. This is for convenience of tracking the ideas.

[abelcarreras]

I agree this part is suboptimal and I didn't pay much attention to this yet.
However, for this optimization 2 degrees of freedom is not enough. You need to sample points of the sphere to find the symmetry elements (for instance rotation axis), but also, at least for some groups, you may have to sample around this axis to find the perpendicular ones and get the complete orientation of the group. That is why I scan the 3 Euler angles (yaw, pitch but also roll).

But I as was saying probably you can reduce this scan to half sphere (and maybe other tricks using symmetry criteria) to reduce the number of points to scan but I didn't think in this part very much yet. If you find a more optimal solution let me know. Using Lebedev grids could be a good solution, I will check your code.