21 maj 2020 kl. 09:21 skrev Daniel Westergren <wes...@gmail.com>: > > Expanding on the discussion about attributes for trails. What's the current > status of the highway=path mess? OSM is increasingly becoming more useful for > forest trails than for car roads (for which other sources are usually more > up-to-date, to be honest). But the default rendering doesn't differentiate > between a forest or mountain path and a paved, combined foot- and cycleway in > an urban environment.
Just as an example, from a data-consumer POV, here's my current rules for rendering highway=path (using the symbols from https://www.lantmateriet.se/globalassets/kartor-och-geografisk-information/kartor/tf_terrangkartan.pdf <https://www.lantmateriet.se/globalassets/kartor-och-geografisk-information/kartor/tf_terrangkartan.pdf>; the Swedish equivalent of the OS Landranger maps). (You can check the results at http://lab3.turepalsson.se/map <http://lab3.turepalsson.se/map>). 1. If it's in a built-up area, and the scale is smaller than 1:12500, don't render it. It just clutters things too much. This rule should probably be moved a bit later to cater for those paths that are more like roads... 2. If it has trail_visibliity = "horrible", "poor" or "no", don't render it. Rationale: If I can't see the damned thing when I'm standing on it, it is useless for navigation. 3. If it has bicycle=designated and (no mtb_scale or mtb_scale < 1) and (no width or width > 2.5), there are some alternatives (rationale: people put bicycle=designated on just about everything, from normal cycle paths to advanced MTB tracks) 3.1 if it has a "hard" surface ("paved", "asphalt", "concrete"), render it as an "approach road, park road, cycle path" (thin solid line). Rationale: This is *probably* a bog standard Swedish cycleway. You could drive a car down it (but you'd have problems passing someone, and it's probably illegal, unless you are a service vehicle from the local council out to empty rubbish bins and change light bulbs). 3.2 If it doesn't have surface, or it has a surface which is not "ground", render it as a "tractor track" (thin dashed line). Rationale: No hard surface, but apparently intended for bikes in some way. The way people usually tag things, it is probably possible to drive tractor, or even a car, if you're careful, but given that we may not have an explicit width, there is a chance of error. 3.3 Otherwise, render it as a "foot path" (dotted line). If we got here, then it has surface=ground, so even if it's bicycle=designated (see 3), you don't want to commute to work on it in nice shoes and trousers. 4. If we got this far, render it as a "foot path" (dotted line). Whatever is is, it has failed all attempts at classing it as anything other than that. What I suppose that I wish to say with all this is that in practice, I have seen highway=path used to mean anything from something that is not even visible on the ground, to something that is impossible to distinguish from a small road (I have seen "single track roads" in Scotland that were on par with some of the highway=path:s I have come across). To know which is is, you have to check several other tags, but you can't rely on any of them being there. OSM tagging is a complete mess of tags describing intention (but not very well -- is "bicycle=designated" a bicycle superhighway or an MTB track?), tags describing legal access (are bicycles required, or even allowed, to use this), and tags describing physical properties such as width and surface and as a data consumer, you have to be prepared for any combination of them. This is my ruleset for highway=path; many of the other highway=* are equally complicated. I'm sorry this ended up so long and unfocused. Map renderer's frustration, I suppose...
_______________________________________________ Tagging mailing list Tagging@openstreetmap.org https://lists.openstreetmap.org/listinfo/tagging