Boyko Bantchev wrote: > Dan, I find your description contradictory > So, what kind of polygons do you mean?
Oof, I was worried this would happen when I wrote my PPS. I wrote: > 2. The polygons are convex. This is correct. > ... if the input is a 5-pointed star, then its central pentagon ... In my case, the input cannot be a 5-pointed star. I just included this thought as a PS (instead of in the body of the email, where the bullets were intended to clarify my situation) because the paper intrigued me with this concept, which would never have occurred to me. I wanted to present my idea of disambiguation via the polygon plurality to see if it floated. Though such a shape can never occur in my use case, if you wanted to generalize the solution, I think it would easy to define an interface which distinguishes between when you want the pentagon included vs excluded. Basically, if you want the pentagon to count as "in", you give all the points of the star at once, and the base edges of the triangles are invisible or are subsumed into the larger shape. If you want the pentagon to count as "out", then give each triangle separately, and so each triangle's base then counts as a legitimate boundary, excluding the pentagon as a consequence. This is a similar reason to why my polygons can't touch or overlap. If I actually had two polygons which touched or overlapped, then their shared or overlapping boundaries would be erased, resulting in a single, larger, polygon. That is, I won't have touching polygons because if I did they would've been combined into single polygon, and presented as such in the first place. -Dan ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
