Matt Sederberg wrote:Thanks for your note. The problem is that models created following strict subd-paradigms are not as smooth when converted to NURBS as those that use T-points. Since star points (which abound in subd modeling) are only g1 smooth, Rhino will not allow us to do anything besides split up the NURBS at every star point.
Yes, I understand that problem, at least at a superficial user-level

.
Matt Sederberg wrote:Also, regarding adding "odd-looking T-junctions", I think this might be an area of user education that we need to do a better job at. I frankly think they look prettier than star points for ending detail since they are smoother!
Yes, when looked at in Rendered Viewmode...
Matt Sederberg wrote:
But, seriously, thanks for voicing your concern--what makes you hesitant to add T-points in your model? Does it somehow make it seem like you have less control over the surface since T-points might still be somewhat foreign?
I was talking about the mesh-flow.
Every Mesh-modelers instinct is to keep a model all Quads, to avoid Ngons and Tris. The result is beautiful to look at. While the the car-model you altered performs way better at conversion time it looks - ugly and "wrong" in comparison, at least when judging from the mesh-topology alone. This is
why I am hesitant to add these modifications manually - and I guess I am not alone with this.
That is why I proposed to perform the necessary mesh-adaptions as a - at least partially automated
process - when converting from imported meshes.
Edit: I should have said "when Converting from Tsplines to Nurbs". I guess it was the easiest for everyone who is trained in SubD-modeling to simply go on working like he/she is used to from
Modo/Silo/Hexagon... whatever SubD-package.