On Mon, 2009-12-07 at 23:22 -0500, Matthias Julius wrote: > IMHO, all those calculations should ideally be done using great > circles. For most practical purposes using linear approximation with > latlon directly is probably good enough for most practical > purposes. Using Mercator for this might give better results.
I think the key point here that the projection used for geometry editing should be a conformal[1] one (i.e., angle preserving) like those in [2]. Possibly, the projection should have an even greater restriction. Matthew W.S. Bell [1] http://en.wikipedia.org/wiki/Conformal_mapping [2] http://www.progonos.com/furuti/MapProj/Dither/ProjConf/projConf.html
signature.asc
Description: This is a digitally signed message part
_______________________________________________ josm-dev mailing list [email protected] http://lists.openstreetmap.org/listinfo/josm-dev
