As discussed in D279, there are some open questions of the BoundingShape implementation such as:
- Why are there always five dimensions of the corrected bounding box?
- Why is the correction based on the minimum and maximum values?
It might be valuable in the long run to refactor the whole class to make it more readable.