CGAL users discussion list

[Narusaki <>]

Hi Sebastien, 

Thank you for the reply. I've found the bug for the extra ray -- it's
because I thought the pleft/pright of the hyperbola mean the left/right
branch, but it turns out it means the left/right part of the branch
according to the axis.

Now it only leaves me with the problem of the shape of the hyperbola. I
gradually multiply each site (x, y, z) by 10^0, 10^1, ..., etc. and the
hyperbola seems to be more and more smooth. I've checked the points
generated by the function you've mentioned, and it's indeed about 100 points
(with an extra one) for whether there is a multiplier for each site or not.
For example, the following two pics are corresponding to multiply the sites
by 10^0 and 10^1 separately:

<http://cgal-discuss.949826.n4.nabble.com/file/n4660484/p1.png> 
<http://cgal-discuss.949826.n4.nabble.com/file/n4660484/p2.png> 

It seems that the first one is approximated by two line segment and the the
second one is approximated by four. I've tried to use the kernel you
mentioned but that doesn't help, and as I've mentioned above, the generated
points number seems to be no problem.

So what the problem could be now?



--
View this message in context: 
http://cgal-discuss.949826.n4.nabble.com/Dual-edge-of-Apollonius-Graph-tp4660456p4660484.html
Sent from the cgal-discuss mailing list archive at Nabble.com.