Hi Folks, > Using transparent overlays is easy - you do not need to do any real time > processing - just get an image from a server and dispaly it. Going for > something cleverer like vector rendering needs a database to give you the > correct data, then the javascript in the web browser needs to process it and > display it. The difficulty is getting a quick enough database response to > give a nice user experience - this was the subject of another 2010 Google > Summer of Code Project > (http://wiki.openstreetmap.org/wiki/Google_Summer_of_Code/2010/AcceptedProjects).
I think client side overlay of features work well to a degree - but ultimately there is more power in full vector rendering. My take on the server responsivitity is that vector rendering clients shouldn't be making full database queeries. I think that a static (periodically refreshed) 'binary tile' architecture where the OSM DB is boiled down into tiles designed for the single purpose of being rendered quickly is the way to go. Ideally a map viewer can then call upon both the standard bitmap tiles and the vector tiles and combine them how it chooses. Cheers Chris _______________________________________________ Talk-GB mailing list [email protected] http://lists.openstreetmap.org/listinfo/talk-gb
