On 24/02/2008 10:53, J.D. Schmidt wrote: > So IMHO it's up to the rendering engines to render the data smartly. > It's not the rendering engines that decide what should be put into the DB.
I'm on both sides here: I agree that it would be better not to have both a node and an area; OTOH, there's already lots and lots of these, so it is shooting ourselves in the foot not to clear up the mess in one way or another. And I'm confused about what Mapnik is actually doing to get this right (if indeed it is always getting it right). There's two ways to do it: either (1) we pass over the data (ideally automatically) and check for parking node within parking polygon, delete the node (or maybe just change its tag, to say parking:redundant, so we don't lose anything accidentally), and transfer any name tag to the area, providing there is no clash,. i.e. the area doesn't already have a (different) name or (2) we change osmrender so it only puts the icon in when there isn't already a relevant node. This would require osmarender to (a) maintain some state: keep a list of parking nodes or areas, and (b) to do the artificial icon after all the nodes (in the layer presumably). Can XSLT do this? In both cases, I think a bounding box test would be sufficient rather than a general point in polygon test, but that's a pretty minor detail. (2) has the advantage that the mapper can override a naive rendering algorithm, and copes with the existing data, but I suspect not knowing much about XSLT that this would be rather hard to do, but it does mean the existing "bad practice" (in quotes because not everyone is of that opinion) is propagated. David _______________________________________________ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk