Bisectors / trisectors / n-sectors for Polygons

[Ethan Ward]

There was a request for this feature (12599). What object should a bisector be? There is an implementation for triangles already, which returns a from vertices to segments from the vertex to the incentere. Would it make sense to just use lines, or rays? What points should be used to define the object, whether it is lines, rays or segments, for a general polygon?

[EKW]

* which returns a dictionary from vertices to segments with segments from the vertex to the incenter

[Chris Smith]

It think the result should be a dictionary with keys being the vertices and the value being the Ray defining the bisector with one point being the vertex. I think that the Triangle implementation should do the same -- it is not necessary to make the result a segment terminating on the side opposite of the vertex (which is an unnecessary addition calculation). An alternative would be to return a Line but it is easy enough to turn a Ray into a Line with `Line(ray)`. The issue you cite gave a suggestion of what to use as the 2nd point of the Ray. I wonder if it might be better to use a point located a distance from the vertex that is equal to the length of the shorter of the two sides that define the angle at the vertex. This has the merit of scaling the Rays in relationship to the vertices from which they emanate.