CGAL users discussion list

[Ashwin Nanjappa <>]

 wrote:
> Hi,
>
> In the current CGAL implementation, there is no flip used for insertion 
> in a 3D Delaunay triangulation.
> The set of conflicting cells (ie cells whose circumsphere contains the 
> new point) is computed, these cells are deleted and the hole is 
> triangulated by simply joining the new point to all facets of the hole.

Thanks Monique. I was looking at the code myself and was able to 
understand those steps. (The hole filling code "create_star_3" though was 
a bit tough to understand due to vague naming of variables like zzz, nnn, 
etc. ;-)

Edelsbrunner's (and Barry Joe's) paper is based on flipping. So, is there 
a paper out there which talks about the CGAL's method of just filling the 
hole which you just described? I'm also interested in understanding the 
proof for CGAL's method.

Thanks,
~ash