These are apparently useful target primitives for subdivision-surface algorithms, that take far less memory than triangle meshes.
It would also make it really easy to have a built-in teapot primitive... :]
Doing this efficiently seems somewhat tricky, but this paper looks interesting:
"A Geometric Algorithm for Ray/Bézier Surfaces Intersection using
Quasi-interpolating Control Net" Yohan D. Fougerolle, Sandrine Lanquetin, Marc Neveu, and Thierry Lauthelier
http://le2i.cnrs.fr/IMG/publications/2188_Fougerolle-RayBezierIntersection.pdf
These are apparently useful target primitives for subdivision-surface algorithms, that take far less memory than triangle meshes.
It would also make it really easy to have a built-in teapot primitive... :]
Doing this efficiently seems somewhat tricky, but this paper looks interesting:
"A Geometric Algorithm for Ray/Bézier Surfaces Intersection using
Quasi-interpolating Control Net" Yohan D. Fougerolle, Sandrine Lanquetin, Marc Neveu, and Thierry Lauthelier
http://le2i.cnrs.fr/IMG/publications/2188_Fougerolle-RayBezierIntersection.pdf