I think it is time to implement 3D machining results in defining 3D cam operations or gears.
I don't want this buried in the manufacturing extension either but it does belong there too.
It shouldn't require a study to show that this can be done "adequately". With rapid prototyping on the rise, we should be able to easily mimic machine tooling operations.
The beginnings of this are already there. The envelope can be united with the trace curve and can be united with surfacing features. At minimum, the operation should be able to generate a smooth envelope surface with edges rather than unusable facet geometry.