* Joseph Eisenberg <joseph.eisenb...@gmail.com> wrote: > Above it was said that the highway=trunk vs highway=primary > distinction is mostly for routing applications. But allowing a proper > rendering is also a main goal of the road tagging system.
> While it's true that road class is useful for routing when there are > two alterate routes, a main reason to tag highways with a certain > class is to be able to render maps properly at different zoom levels. > When you are making a high-scale, low-zoom-level map of a large area > (say, the whole State of Alaska, all of England, or all of Australia), > you will want to only render highway=motorway + highway=trunk, because > showing all highway=primary would lead to rendering many smaller roads > which are not reasonable to show at that scale in most places. > [..] What makes the problem of road classification so hard is that we want it to do different things at once. For rendering we have on the one hand the requirement that we want to show all the "relevant" roads for a given zoom level, on the other hand, as a map user I would expect that a road shown as e.g. trunk in Massachusets would be quite similar in characteristics to a road shown as trunk in Montana. I suggest to keep the road classification consistent at least within a country and try to solve the problem of roads in low-zoom maps at the rendering level, by modifying the list of displayed road classes until a target density of displayed roads is reached. That might become easier to do when we move to vector tiles. Wolfgang ( lyx@osm ) _______________________________________________ Tagging mailing list Tagging@openstreetmap.org https://lists.openstreetmap.org/listinfo/tagging