Would it be ok to determine if a stroke is a circle by seeing if each coordinate has a matching coordinate directly across from it? This would distinguish a circle from a line, but would not distinguish a circle from a square. You should take a pragmatic approach to questions like this: if it works it's ok; if it doesn't work then it's not. The difficulty I see (after thinking about it for all of a minute or so) is trying to figure out what "directly across" means. Another difficulty I see (if I understand you correctly) is that you may be begging the question; "directly across" make sense with respect to circles but not lines, so you have to determine if a stroke is a circle or a line before you can apply your test to see if the stroke is a circle or a line. I don't want to discourage you, it may turn out to be a good approach; you've thought about it longer than I have. But if you can't handle these two difficulties, the approach probably isn't fruitful. I will say there are simpler approaches to distinguishing circles from lines. And remember, you have to distinguish circle-like strokes from line-like strokes, you don't have to distinguish between different kinds of circle-like strokes.Received on Thu Sep 20 2007 - 14:50:22 EDT
This archive was generated by hypermail 2.2.0 : Mon Sep 24 2007 - 05:35:52 EDT