Propose the fix for #522.

[kellertuer]

No description provided.

[kellertuer]

Maybe an alternative to have both variants as a possibility would be to keep the function as is and do the Lie algebra projection for the group manifold analoga.

[kellertuer]

I feel this variant should actually be the best choice: To have both.

For _Manigolds (Rotations, OrthogonalMatrices, UnitaryMatrices) we keep the representation as is (that is X is the tangent vector and not necessarily skew, but a rotated skew symmetric), for Lie groups (Orthogonal, SpecialOrthogonal, SpecialUnitary, Unitary) we use the Lie group representation. This way it is easy to have either of the representations, depending on what the user wants – in a precisely distinct way.

[kellertuer]

hm, seems really more complicated than I thought. So I will wait for the decision whether we want to unify it and in which way.

[kellertuer]

So if I understand correctly it is not only complicated to finish this but even more to get it work together with differentiation again? I would maybe propose to close this?

[mateuszbaran]

This would be a lot of work to finish for not much benefit so I'd suggest closing it.