Geometry: store references to CoordinateSystem in low-level vs. high-level objects

Currently, all low-level geometry objects (Point and Vector) have their own pointers to their corresponding CoordinateSystem in which they are defined. High-level objects such as the Helix have multiple Vectors and Points as members. Therefore, there can be a lot of redundant information stored in the high-level object if all members are in the same CoordinateSystem and the members can even be defined in different CoordinateSystems, which I consider undesirable.

Are we ok with that or should we devise a better implementation? I have a vague idea…