[OSM-dev] Slippy map way forward -- OpenLayers and/or Mapstraction?

nick at hogweed.org nick at hogweed.org
Fri Jun 30 09:22:04 BST 2006


>I was wrong. http://openlayers.org/dev/examples/openstreetmap.html is an
>example of that, but it clearly doesn't work. Roads which clearly cross
>tile boundaries dissappear when they hit the boundary, roads are drawn
>right through the millenium dome, etc. 

What happens at the moment with the OSM slippy map and renderer (and what has 
always happened, I think) is as follows:

a) OSM slippy map takes in a bounding box in normal (non-Mercator projected) 
latitude and longitude
b) Internally (in the slippy map), the latitude is coverted to 
Mercator-projected latitude in order to divide up the tiles evenly by 
Mercator projection, so that the tiles won't be "squashed" by the poles
c) For each tile, the Mercator-projected latitude is converted back to normal 
latitude and sent to the OSM renderer. This is because the OSM renderer 
accepts a bounding box of "normal" (WGS84? not Mercator-projected anyhow) 
latitude and longitude. (I thought that was standard? i.e. I thought WMS 
servers accepted a bbox of regular latitudes and longitudes, not 
Mercator-projected values)

If this is incorrect, it would be useful to know so that the code can be 
changed as necessary.

Nick




More information about the dev mailing list