got a problem with the accuracy of NURBS. I wrote a viewer for NURBS based models. The problem is: On which value should I set the GLU_SAMPLING_TOLERANCE? I don’t understand this value, in the docu they describe it as a pixel based value.
When I set this value e.g. to 550.0 it is ok with one model. With another model I need to set it to 550000.0, otherwise opengl will eat 500 Megs of memory!
Any hints how I can calculate a good value depending on the model size or something else?