Rotations in 3D

[hdavid16]

Current implementation uses tangent angles and Billboard for rotations.
I noticed the comment in to_angle: For some perspectives this might give wrong angles in 3d

Is this because we are using Billboard, which takes an angle? Should be we using Quaternions for 3D space?

[hdavid16]

Actually, 3D rotations do work in with Billboard. So I guess no need to do this....?

[hexaeder]

I think the problem is that we need to figure out the point t where the arrow is close to the node marker in projected pixel space rather than data space. In 2d no problem because its just a linear mapping from data to pixel space. In 3d its more complicated. I think quaternions would be neede in case of 3d arrow 'cones' rather than flat sprites like the makie arrows plot do.

[hdavid16]

Ahhh yes. Makes sense.