hi,
a question about ‘boundary surface’ component:
One would expect it to create a planar surface, aligned with the world axes and then trim it down to the perimeter of the curve.
Very curiously though I noticed that often, the surface is aligned to the curve (e.g. when the curve is rotated around the z axis, the surface is constructed with that orientation.)
Generaly it is not important, but I’ve found quite a few times that the orientation of the untrimmed surface is important further down the pipeline.
Does the curve carry some kind of metadata? (like orientation/local axes or something?). Any insight would be appreciated!
No, but the curve has a shape that can be evaluated, and when converted into nurbs, it has control points that give a limited set of possible orientations for which the aligned boundingbox would be minimal. Or, it allows us to find a rectangle which shares an edge with the curve, thus requiring no trimming on that side.
The less area of a surface is trimmed away, the more efficient that shape is. The function in Rhino which creates these surfaces assumes people do not care about UV orientation. If you do care, I do not know how to easily change that. For single curves yes, you can just start with a big enough oriented surface and split it, but when you have nested curves it will be difficult to convert the inside/outside logic to a fresh split.
Thanks! that answers it completely!
I have a few definitions that are based on my false assumption that the surface is always parallel to the world. Nothing that can’t be solved with a BBox.
Sorry If I’m being pedantic.
Given the opportunity [I have a feeling that making the right regions from closed curves is more important for your type of work], get some regions related stuff for the cold winter nights (or the warm ones if you listen to The Lord).
