At the risk of getting in a battle with you Albatros, I will respond to a few of your misleading statements:
"The fact that C.C. allows the user, either intentionally or not, to modify/change the original concept without regard to its original perception, sets up a situation when a blueprint/drawing may have edges, faces, curves and confusing lines unaccounted for and not related to the real geometry."
This is obviously written by someone who has not used CoCreate. Firstly "to modify/change the original concept without regard to its original perception" is often exactly what you want to do in the conceptual design phase or when re-purposing an existing model. Being tied to the constraint hierarchy that you (or somebody else) created to model something is usually a hinderance when you want to make drastic changes which is often the case early in the design, when you starting with someone else's models, or when you want to re-purpose a design into something new. This freedom is precisely what is great about direct modeling.
Secondly, the idea that with CoCreate this "sets up a situation when a blueprint/drawing may have edges, faces, curves and confusing lines unaccounted for and not related to the real geometry" is not true. When you make your 2D drawing, it will only generate correct views based on valid 3D models. If you turn a model into a face part (intentionally or not), it will not allow your drawing views to be refreshed. You have to either correct the model (turn it back into a solid), or specifically and willfully modify the view to allow calculation of face parts. This is sometimes desirable and the results are just as accurate as views created from solid models, you just have to think of the face parts as made of sheets in space and what the resulting 2D views would look like. That is what you get.
"One reason why constrains exist, we may argue about it, is to show what the limitations are and how to prevent user screw-up’s"
I don't think it was ever the intent of parametric modeling to prevent screw-ups. It is a different philosophy that allows the capture of design intent and provides for expressing variations of a design. For mature and production designs, it often makes sense to use parametric modeling. It is quite impressive the way that a large assembly can handle small changes buried deep in the hierarchy to update adjoining models, assemblies and drawings. This process is necessarily more manual and less efficient with the direct modeling method. This is the liability to the freedom you are granted in creating and modifying the models using direct methods. I think a better way to look at it as others have said is: "which tool is best for the situation at hand?"
As I stated in an earlier post to this thread: "So I think the direct method is better for the conceptual design phase and/or working on other peoples models. The constrained method is better for mature designs and production designs were the up front effort of capturing the design intent pays off in the ability to handle the rippling effects of small variations on the design."
So maybe a better way to recast the thread title is: Will a Pro/E or CoCreate user ever be able to use either parametric or direct methods as needed when it is the most efficient tool for any given modeling task?
