Subject: Re: [geometry] Segment to Segment Distance
From: Mats Taraldsvik (mats.taraldsvik_at_[hidden])
Date: 2012-05-25 10:04:28
On 05/25/2012 03:40 PM, Arash Partow wrote:
> On 25/05/2012 9:57 PM, Barend Gehrels wrote:
>> But is this related to Boost.Geometry? Or even Boost? No.
>> Please consider the context of this list.
> Sour grapes? ;D
> Based on the last paragraph of your reply (does it really need to be
> that complicated?) and the original poster's need for a simple
> solution, I thought I'd help out.
Boost.Geometry does allow integration with custom, existing types, e.g.
for integration with legacy classes/objects. The higher flexibility does
make the code more complicated, but the savings in runtime performance /
memory makes up for this, imho.
Other libraries, like the one you advertised (it is your own library,
after all :) ), does not have this flexibility (correct me if I'm wrong
here, I took a quick peek at the links) -- i.e. if you want to run your
wykobi::algorithm::sutherland_hodgman_polygon_clipper, you first need to
create objects of the class wykobi::rectangle and wykobi::polygon, which
comes at a cost in memory and cpu. With Boost.Geometry, you can use an
arbitrary class for the algorithms, as long as you adapt e.g. the point
or polygon concept.
> btw using ggl these days to do computational geometry is like
> attending a poncy cheese tasting afloat a sinking barge sailing down
> 'The Sarno' ;>
Since the features of the libraries, and their intention, are different,
the comparison is moot. Also, remarks like this are unnecessary.
> Geometry mailing list
Geometry list run by mateusz at loskot.net