rom: 
          Peters, Tom
                                                         Sent: 
                                                                     Wed 11/10/2004 4:17 PM
 To: 
          Peters, Tom; Alana Tuccillo; eric.borbely@uconn.edu;
         Michael.Crivello@uconn.edu; Rina.Shah@uconn.edu;
         CARLETON@COFFRIN.COM; mariusz_gola@comcast.net;
         Nicholas.J.Thompson@uconn.edu; DML1001@yahoo.com;
         Chris.Puzzo@gmail.com; daniel.Fontaine@gmail.com;
         Toan.2.Nguyen-Dinh@uconn.edu; thaibinh.luong@uconn.edu; Yang, Zhi;
         nguyengochuy@hotmail.com; yanroy@comcast.net; Jeremy.Spitz@uconn.edu;
         QueenBeee777@yahoo.com; joshua.levinson@uconn.edu;
         bjorgejm@hotmail.com; Rina.Shah@huskymail.uconn.edu
 Cc: 
          grendelkhan@gmail.com
 Subject: 
          CSE 275 Point Inside Tests

 Attachments: 
                                                                       View As Web Page

 The class of 11/11/04 will focus upon the following:

 Point Inside (Polygon and More Generally)

 Ray casting versus winding number:

 Map each point of curve to circle via unit vector.


 http://softsurfer.com/Archive/algorithm_0103/algorithm_0103.htm#Winding%20Number

 Still may need to pick line that doesn't hit any vertex or edge, but
 we have a randomized strategy for that.

 Seems to suggest should exclude horizontal edges and make test with
 horizontal ray, but what if grazing in floating point.

 Also, equality of vertices in floating point.



 http://www.acm.org/pubs/tog/editors/erich/ptinpoly/

 He trashes winding number approach, as terrible because it is so slow, but
 what about being more conservative for correctness?