Re: [boost] [geometry] [impl] polygon relations [was: area area, length, centroid, behavior]

Subject: Re: [boost] [geometry] [impl] polygon relations [was: area area, length, centroid, behavior]
From: Simonson, Lucanus J (lucanus.j.simonson_at_[hidden])
Date: 2009-03-13 13:17:09

Phil Endecott wrote:
> Barend Gehrels wrote:
>> http://www.cs.man.ac.uk/~toby/alan/software/
> 
> I just thought I ought to mention that that code is NOT FREE, so you
> may want to "talk to your lawyer" before looking at it.  (I'm aware of
> one application that removed that code during a license audit; I was
> wondering if any future Boost geometry code could end up replacing it
> there.) 
> 
I think we are safe to benchmark it, writing an academic paper is non-commercial use, I don't plan on reading their code.  If they haven't written any papers (I've never seen any from them) I may choose not to discuss them at all other than as a curve in a runtime comparison graph.

I certainly hope a future boost geometry code could end up replacing GPC as a good free alternative.  I believe what I currently have is good enough to be a drop in replacement.  I provide a superset of their features, I am probably no worse in runtime even with my sub-optimal line intersection or arbitrary angle geometries and I am 100% robust for the full integer coordinate domain and all degenerate inputs.

I'll be updating what is in the sandbox today, if you would like to explore the applicability.  I'll check in my unit testing cpp file, which can serve as example code for how to use just about everything in the library since the purpose of that thing is to get my code coverage up and verify correctness of everything I write.

Luke