Re: [Boost-users] [Polygon] Test among 3 libraries

Subject: Re: [Boost-users] [Polygon] Test among 3 libraries
From: Angus Johnson (angus_at_[hidden])
Date: 2010-09-08 09:44:19

  Simonson, Lucanus J 
<http://search.gmane.org/?author=Simonson%2C+Lucanus+J&sort=date> wrote:
"I've never seen a comparison with clipper before, I've never heard of 
clipper and would like to learn more. Can you provide a link or 
reference? I found a sourceforge site with minimal description of 
clipper, but it left me with a lot of unanswered questions."

Hi Luke.
I'm the author of Clipper and am happy to answer your questions as best 
I can, but perhaps I can preempt some of them with a little background 
about Clipper and myself.

Firstly, I've only recently created the Clipper library (ie within the 
last 3-4 months) and that may explain why you haven't heard of it. I 
originally wrote it in Delphi Pascal, but I've translated it into C++ 
too, and has been released with a Boost license. I wrote Clipper because 
I wanted a polygon clipping library (in Delphi) that wasn't encumbered 
with restrictive licensing. I'm still refining the library and have 
several updates in the pipeline (including polyline-polygon and 
polybezier-polygon clipping).

Clipper is based on the Vatti clipping algorithm which places no 
restrictions on the types of input (subject and clipping) polygons. 
However, Vatti's original paper completely omits any discussion of 1. 
how to handle what I call 'complex' intersections - intersections 
involving more than 2 edges; and 2. how to handle overlapping horizontal 
edges. My Clipper library extends the Vatti algorithm by addressing both 
these issues. Clipper also extends the Vatti algorithm by implementing 
the NonZero polygon filling rule for input polygons (as well as Vatti's 
EvenOdd filling). I'm aware of one limitation of my Clipper library: 
because it uses floating point math it isn't 'robust' in the strict 
sense. However, errors (which are extremely rare) are properly trapped 
and reported as a failed clipping operation.

In my testing ( see http://www.angusj.com/delphi/clipper.php#features ), 
Clipper has a number of advantages over GPC, a well known clipping 
library that's also based on Vatti's algorithm. Clipper - 1. is 
significantly faster; 2. properly handles complex intersections; 3. 
accepts both EvenOdd and NonZero filling input polygons; 4. doesn't 
require payment for commercial use.

The SourceForge site for Clipper is: 
http://sourceforge.net/projects/polyclipping/

About myself: I've been writing computer code for 30yrs as a hobby. (I 
was a medical practitioner before I retired 15 yrs ago due to illness.) 
For the last 15 years I've been using Delphi (Object Pascal) and until 
now have never used C++. The C++ translation of my Delphi Clipper 
library is the first code I've written in C++ and should explain any 
unorthodoxy in my C++ coding style. (I'm aware that the C++ code is 
currently about 15% slower that the Delphi code so I suspect I'm not 
making the most of this language.) I'm very happy to receive 
constructive feedback and hope I can answer any questions too.

Angus