(Rhino V7 / V8, Rhinocommon, c# plug-in and scripting)
I am afraid I miss something here - how to create a surface with minimum of inputs / spans / CV’s
A simple example that illustrates my question would be a custom variable fillet surface.
Let s say my algorithm will give any desired / unlimited amount of sections, meaning for points on both surfaces and Arcs / shapes.
My current approach to create a surface is to increase this amount of sections until the result of loft or sweep2 is within document tolerance. I increase the overall number of sections.
Rhinos surfaces (sweep2, filletSrf, …) seam to use a different, more efficient approach.
As an example:
left Rhino s VariableFilletSrf
right my brute force approach with to much shapes/spans/CV’s
@menno or @dale
Do those commands use the logic of sweep2 under the hood?
Any relation between creating the rails and the shapes
Any rules to follow, for increasing the number of shapes locally ?
Variable radius FilletSrf uses adaptive fitting to get the radii within tolerance along the surface edges. This is why you see some regions with denser control point lines at the very left and towards the right of the result.
Very basically this is 1) getting radii as a function of distance along the common edge followed by 2) start with a surface for a few radii, and test if radii in between those used are within tolerance. If the tested radii are not within tolerance, those are added to the list of radii to use for surface interpolation and the surface is re-create with the added radii. Keep doing this while any tested radii are not within tolerance.
