Re: [boost] [Review Request] Inclusion of the Boost.Polygon Voronoi Library

Subject: Re: [boost] [Review Request] Inclusion of the Boost.Polygon Voronoi Library
From: Vadim Stadnik (vadimstdk_at_[hidden])
Date: 2012-04-11 07:43:35

Hi Andrii,

The algorithm is useful, no doubt.

The question concerning fully constructed Voronoi diagram as well as
Delaunay triangulation:

What is the computational cost of operations for a single element, such as,
find the nearest neighbor, insert/erase a vertex?

How long, do you think, it can take to parameterize types of coordinates,
points and point predicates?

The benefits of the generalized algorithm are quite significant. Many
existing libraries can use this algorithm either directly or with minimal
adaptation. The parameterization of point predicates is quite important.
For example, in practice, tolerance based predicates are very good
alternative to multi and adaptive precision methods. With the right choice
of predicates points using floating data types will be supported too.

Regards,

Vadim Stadnik