How does Rhino determine whether a curve is an arc or part of an ellipse?

Replies to This Discussion

Permalink Reply by David Rutten on October 20, 2016 at 3:26pm

Can it tell whether it's elliptic?

I imagine three points are chosen on the original curve, a circle is fitted through these points, and then additional points on the curve are tested to see if they are near enough to the guide circle.

Permalink Reply by Ethan Gross on October 20, 2016 at 3:42pm

Well, if I shatter an ellipse, GH tells says the fragments are still elliptical curves. Maybe Rhino's just assuming a shattered ellipse or circle is still composed of similar curves and doesn't bother recalculating.

Permalink Reply by Mateusz Zwierzycki on October 20, 2016 at 3:46pm

"Elliptical Curve" doesn't mean it's an Ellipse or a Circle. It can be, but it can be a part of it as well.

Permalink Reply by Ethan Gross on October 20, 2016 at 3:51pm

This just brings me back to my original question, how does Rhino "know"?

Permalink Reply by Pieter Segeren on October 21, 2016 at 3:44am

Could it be Rhino tries several "Tryget..." methods on the curve? There are for instance TryGetArc and TryGetEllipse methods (among others) for curves.

Permalink Reply by David Rutten on October 21, 2016 at 3:50am

I don't know, and I never seem to be able to find answers to these kind of questions by digging through the Rhino source. I'm afraid you'll have to ask on the mcneel discourse forum where the Rhino programmers hang out.

Permalink Reply by Ethan Gross on October 21, 2016 at 5:54am

The funny thing is, you can rebuild a circle or ellipse with enough points, regardless of degree, to resemble the original at practically an atomic level and it will still just reference as a "periodic planar curve", so they must be specified in some unique nurbistic manner.