On Mon, Jan 31, 2011 at 9:44 PM, M∡rtin Koppenhoefer <dieterdre...@gmail.com> wrote: > Yes, parallel ways are actually to be considered errors in the case > that the polygons really do touch, while colinear ways are simply less > elegant but not wrong. I fear though, that laziness will lead to > people identifying features as touching, when they actually aren't in > the real world (e.g. between 2 fields there is often some rough area, > sometimes with a waterway running along, and personally I prefer > mapping these fine details, even though it is more work then > approximating the situation).
I think a few of our tags aren't sufficiently well defined to be clear about which is correct. For example, say a steel manufacturer owns a large piece of land, at one end of which is a steel smelter. The rest is grass. Is the whole land landuse=industrial? Just the built-up section? Does the landuse correspond to the physical features or the zoning... Partly this gets back to the debate about whether a map is an abstract representation, or whether it's essentially a 2D bitmap of the real world. But...perhaps let's not go there right now. :) Anyway, the range of opinions in this thread convinces me that there is no consensus yet about when and how colinear ways, parallel ways, and multipolygons are to be used. Improved tools for all three should at least help people use whatever is the preferred tool for the job, rather than being constrained by whatever is the easiest. Steve _______________________________________________ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk