Subject: [ggl] Proposal for Delauney triangulation algorithm
From: Barend Gehrels (barend)
Date: 2011-10-29 16:34:44
On 28-10-2011 2:28, John Swensen wrote:
> That being said, I have a few questions:
> 1) What is the appropriate way to split a polygon? In the detail portion of the algorithm, in the specialization for polygon, it would be easiest to define portions of the Ranges for the exterior and interior ranges to be combined to make a new polygon. Maybe this isn't possible, but would be interested in an example of this if one exists.
Without diving into the algorithm, this is hard to answer for me. IMO
you have to take both exterior and interior rings into account, so
cannot process then independantly (which most algorithms do... even
intersections, more or less).
> 2) Return types for algorithms. Is the approach taken in the union_ algorithm the correct approach to get one geometry in and another geometry (or better said a multi_geometry) out?
Yes, this is the way how we do it. We don't want to "return" it so
specify it as a parameter. The intersection and union_ take a vector of
polygons. Users may use either a multi_polygon there, or a vector of
Geometry list run by mateusz at loskot.net