generative modeling for Rhino
Well the question is in the title. Is there a way to test for this property, other than sampling and per-maybe-haps using isovist? the trouble is in the sampling part you can imagine. I have not tried to tackle this yet, so I'm just throwing the question in...
Huh sorry I meant like this...
I should have said convex and non-convex to be precise - but I did say closed, didn't I? ;)
Rough definition of convex : if you have a region (2d or 3d), take two arbitrary points within it and draw a line. If the "shape" is convex the the line will never cross the region boundary.
so a sphere or a cone are convex, a torus isn't.
Thanks for your input Dany, always first on deck to help people in trouble. But I'm not a beginner really! I'm thinking actually I might move this post to the vb forum. It must have been implemented in some open library. I hope...
I missed the closed :) no worries.
There's nothing ready made for this. One interesting approach could involve Convex hull. If a curve is purely convex, it will be similar to the convex hull of all the points that make up the curve. Also if a curve is convex it means its control-point-polygon will also be convex. Comparing the control-point-polygon to its own convex hull may give you the answer.
Yes you're right, I'm gonna look into this. Thank you David.
On second thought, are you sure that the opposite is true? If the control points polygon is non-convex then the curve is non-convex too? and what about CP weights?
Yeah you're right, it is possible for the control-point-polygon to be concave while the curve remains convex.
Thank you for this one fred,
exactly what I was after.
(for the definition I am using, this results (+ to - change)
show me if the curve remains convex or not)