Am 02.01.2014 17:39, schrieb Fernando Trebien: > I'm quite convinced that it's impossible to pick only one tag or the > other for this particular rendering decision and please everyone at > the same time. That's why I'm still in favour of combining them, then > each community can use whatever they like, and the discussion of which > one is the best would become solely a tagging concern. The renderer > would (like many other apps) simply abstain from influencing mappers. > Does anybody disagree? > > If nobody disagrees, I think we should only decide whether the > following tag combinations are to be considered paved or unpaved: > - tracktype=grade1 + surface=unpaved/ground/earth/dirt > - tracktype=grade2 + surface=paved/asphalt/concrete/paving_stones > > Or, in other words, which tag takes precedence when deciding the way's > "paved" status (which is purely a semantic problem - so please forget > about the renderer for one moment). I think surface should take > precedence, as it states that idea more clearly. Is this a matter of tagging, which takes precedence in the decision a router makes?
Tracktype and surface may be similar or orthogonal, both is possible. I know (without being able to show you photos or something like that) ways that are paved with paving stones (and thus clearly counted as paved), but due to tree roots below the way and so on are tracktype=grade3 or worse; and on the other hand a well maintained compacted road may be grade1 in certain weather conditions (and certain regions where the weather is "stable" in a good condition). Therefore I think, something like "quality", partly encoded in tracktype currrently is clearly distinct from surface, although it sometimes correlates. If we fix a derived meaning to BE USED by routers and TO BE KNOWN to mappers, we lose the meaning of the single tags, and one tag routing_quality=1..10 would suffice (while not easily to understand). For routing using surface alone may be a good decision in some parts of the world, and a bad one in others; same goes for tracktype alone. A combination of both may be arbitrary, and a dedicated new track too. The solution for routing may be to use different routing profiles for different parts of the world or different preprocessing. regards Peter > > On Thu, Jan 2, 2014 at 2:15 PM, gweber <gwebe...@gmail.com> wrote: >> Sorry for coming late to this discussion. I have been trying to get some >> attention to the surface=unpaved rendering issue for some time. Fernando >> alerted me that this discussion is now being held in this discussion list. >> >> Essentially the problem is that currently there is no way to alert the users >> about unsurfaced roads. In countries, such as Brazil, this can easily become >> a dangerous situation. >> >> Please have a look at this location http://osm.org/go/Nq9hwSPZ >> <http://http://osm.org/go/Nq9hwSPZ> >> >> One of the tertiary roads is halfway unpaved. Can you see where? No one can >> unless you switch to HOT style. Even then the rendering difference is too >> subtle to serve as a warning. One way to work around this deficiency would >> be to use highway=track and indeed that is what many people do. However this >> is clearly the wrong approach. >> >> I would strongly favour a simple dashed border style whenever the surface >> tag falls into the unpaved categories. It is as simple as that. From my >> experience in driving on rural roads in Brazil, nothing else is required. >> >> I've seen the discussion about using tracktype for this purpose. I think >> that would be ill advised as it would generate yet another semantic >> inconsistency to OSM (using track-something for non-tracks, very much like >> classifying highways as unclassified). So please, don't. >> >> I agree that the surface tag is not be as widely used as desirable, but that >> may change once it is being rendered. >> >> Finally, I can assure you, from my own driving experience, that >> surface=compacted should be considered unpaved for all practical purposes. >> >> >> >> >> >> >> >> >> >> -- >> View this message in context: >> http://gis.19327.n5.nabble.com/Tags-useful-for-rendering-of-roads-in-poor-conditions-tp5791303p5791478.html >> Sent from the Tagging mailing list archive at Nabble.com. >> >> _______________________________________________ >> Tagging mailing list >> Tagging@openstreetmap.org >> https://lists.openstreetmap.org/listinfo/tagging > > > _______________________________________________ Tagging mailing list Tagging@openstreetmap.org https://lists.openstreetmap.org/listinfo/tagging
