CGAL users discussion list

Wesley Smith wrote:
> You can look at the 3D Delaunay class and use the dual() functions to
> figure out the voronoi structures. 

The Voronoi combinatorial structure is explicitely stored in the 
Delaunay triangulation (up to degenerate cases). There was a thread on 
the same topic a few months ago (I would say august).
The geometric embedding (= coordinates) of the Voronoi diagram can be 
computed by the dual() functions.

> Unfortunately, the dual of a
> Delaunay vertex in 3D is not available as a function so you have to
> write it yourself. 

We could probably provide a Polyhedron_3 as the result of the dual() of 
a Delaunay site. I am not sure it would be very useful.
The vertices of the Voronoi cell of a site can be accessed by using the 
iterator on its incident cells in the triangulation and the dual() on 
these cells. Facets of the Voronoi cell can be accessed from the same 
iterator.

> I've done quite a bit of work toward making an
> analog of the 2D voronoi class in 3D but the code is not ready for
> others to use. 

It would be nice to have such an interface class in CGAL.

best
        Monique Teillaud