Apologies, I hadn't noticed this part of your post earlier, I will address it here.
Actually, almost nothing about 3D objects was changed in any way to accommodate the screen vs layer plane system. This was due to the current way features are selected, tested and implemented. To you and I, it is very obvious that altering the way 2D and 3D objects are displayed in a given "mode" will dramatically affect most if not all of a hybrid modelling process, but this was not noticed during the implementation of the task.
It wasn't an intentional decision to make parts of that system confusing or clunky, just a side effect of focusing on some smaller portions of how it was implemented while the implementing engineers were unaware of others/the larger picture. This is something I would like to see corrected in the future, however it will take some time and effort, I only have so much pull but I plan to use all of it.