On 20-Nov-2009, at 10:52, Kris Douglas wrote: > 2009/11/20 Xan <[email protected]>: > >> @James: if google did it, it's useful. The enterprises don't waste >> time.... They want to make money. So that information is useful in some way. >> > > As a matter of fact, Google did do it for fun. They had an existing > system for Google maps, and decided to use free time to throw some > basic map data on to the engine. With the google engine, I could do it > in 20mins with the data.
Having looked at how things are done, the main challenge would be mostly with keeping the bandwidth down, than anything else. The server is essentially serving up tiles of varying sizes based on the zoom and the front-end is essentially deciding which tiles to fetch based on what is currently displayed and in which direction the user is moving the map. I was able to split one of the higher resolution maps from NASA using ImageMagick, which gives the tiles. I then made a basic UI, which I never got round to finishing that would fetch the image based on its name. For example: /..../<zoom-level>/image_<x>_<y>.jpg The code would know about how many tiles high and wide the map was and also the size of the tile for the given zoom. I haven't looked at how OSM draws its maps, but there is a good chance it may be doing the same thing. André-John _______________________________________________ newbies mailing list [email protected] http://lists.openstreetmap.org/listinfo/newbies
