Dirk-Lüder Kreie <[EMAIL PROTECTED]> writes: > Maarten Deen schrieb: >> I'm a bit puzzled for the reason why this gets rendered as it does: >> <http://informationfreeway.org/?lat=-84.951&lon=-144.66&zoom=12&layers=B0000F000F> >> >> Tiles are >> <http://server.tah.openstreetmap.org/Browse/details/tile/12/402/4083/> >> <http://server.tah.openstreetmap.org/Browse/details/tile/12/402/4082/> >> <http://server.tah.openstreetmap.org/Browse/details/tile/12/401/4082/> >> >> In JOSM, this is a straight line. Why does the render mess this up so much? >> Two tiles have a turn in the coastline, and the middle tile is not >> in line with the outer two. >> >> If you look in JOSM with the slippymap plugin, the rendering is way >> off of where it should be. >> >> And the coast is riddled with these tiles. >> Is this some side-effect from being close to the pole? > > I suggest the following possible causes (non exhaustive) > - error in beziercurving > - stale datasource (way was moved, ROMA/XAPI did not catch up) > - rounding error when projecting > - general projection error in one of the clients > - coastline had added nodes only every second tile, so those got > updated (autorequest), while the others kept the old state of affairs > (no autorequest due to no node change), which was apparently a bit > different.
But it is noticable that the coastline is relatively higher if it is near the bottom of the tile and lower if it is near the top of the tile. And this is pretty consistent. So it seems to be a systematic error. Matthias _______________________________________________ Tilesathome mailing list [email protected] http://lists.openstreetmap.org/listinfo/tilesathome
