Michael, Thanks for the wisdom on this subject. It was one that I've contemplated at times.

You'll have to forgive my ignorance on NURBS dynamics... ;-)
But perhaps, could a hybrid solution could be made?

Not a new kind of patch, but a method where both methods are first created: the 'single' patch and the 'star'. And the best of both are used.

Actually, the center region of the single patch seems (relatively) smooth enough, perhaps the resultant control points in the 'star' version's center vertex could be - to some gradated degree - conformed to the surface positions of the 'single' patch's center?

Or in other words, couldn't the relatively consistent surface distribution of the center area of a single patch be used to tightly conform the controls of the star-type's vertex area?




I have a feeling that the underlying weighting and strange complexity of surface creation are the real culprits here.

Maybe not?... I tried. ;-)