[OSM-dev] Algorithm help

Richard Fairhurst richard at systemeD.net
Thu May 29 02:12:52 BST 2008

Thanks Richard and Nathan for the pointers - really appreciated. I  
think I've got it now.

Nathan Vander Wilt wrote:

> [...]
> That said, I'm guessing there are other assumptions being made  
> within OSM that break down under spherical geometry. If your polygon  
> doesn't contain pole(s) or cross the +/-180 degree meridian,  
> checking the winding could be handled by a cartesian algorithm.

Yep - it's so that Potlatch can show the direction of a closed way;  
happily Potlatch works with projected co-ordinates anyway and is  
(happily) unlikely to have to load a way greater than a hemisphere. :)


More information about the dev mailing list