From: schikore AT kavala DOT ccad DOT uiowa DOT edu (Matt Schikore)
Newsgroups: comp.os.msdos.djgpp
Subject: Re: bounding circle vs. bounding boxes
Date: 26 Feb 1997 22:05:17 GMT
Organization: University of Iowa
Lines: 31
Message-ID: <5f2c2t$uvs@flood.weeg.uiowa.edu>
References: <19970226 DOT 065019 DOT 4511 DOT 1 DOT fwec AT juno DOT com> <33146FD7 DOT A93 AT pobox DOT oleane DOT com>
NNTP-Posting-Host: kavala.ccad.uiowa.edu
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp

>The best (ie smallest) bounding circle can be calculated as follow : 
>find the two points A and B in your set which are farthest from each 
>other, the best circle has the middle of segment AB as its center, and 
>half the distance between A and B as its radius. 

This doesn't seem like it would work.  If I have the following 
triangle ABC:

                  A
                 / \
                /   \
               /     \
              P       \
             /         \
            /           \
           /             \
          B---------------C

All sides are equal.  So we take side AB to be the longest.  By your
algorithm, P would be the center of the bounding circle, and the radius
would be half the length of AB.  This circle would obviously not contain
the point C.

For this specific case, the best bounding circle is obviously centered
at the center of the triangle, not on one of the edges.

-Matt Schikore
schikore AT ccad DOT uiowa DOT edu

--