2009/9/16 Lennard <[email protected]>: > I commented on the ticket. This is a better place to discuss this than on a bug tracker:
> This aussie patch is /an/ approach, but not necessarily the best. It makes > relatively expensive calls to postgis to get the Aussie boundary and compare > each route to that. While that may look innocent enough, try extrapolating > doing this for 50+ countries. That's even discounting the explosion in the > size of osm.xml. >As a way to handle the expensive postgis calls: I think a much better approach >would be to determine locality of features during import, and not during >rendering. The import will take longer, sure, but tile rendering should be >much faster than using your run-time filtering method. > On the matter of stylesheet maintainability, XML entities definately come in > very handy here, but it will still be a humongous stylesheet. Other > approaches like cascadenik should not be discounted for this. I didn't think it would be the best option, for most of the reasons you point out, however we have to start some where, so I thought I'd get the ball rolling. _______________________________________________ talk mailing list [email protected] http://lists.openstreetmap.org/listinfo/talk
