Andy Allan wrote: > highway tag with more than 100 instances. You can see that the > highway=footway/cycleway/bridleway/pedestrian totals 509,920 instances > (path has a respectible, but tiny by comparision, 2165). And everyone
Meh, bridleway/cycleway/footway have been around for at least 2 years, path's been around for what, a few months? Pedestrian's not in the same class at all though, so make that 425756. Not that it reduces your point much. > of renders, routing algorthims and whatnot understand and use them > too. Routing algorithms need to understand the access tags anyway, so no real change there. Renderers should as well (without bringing path into it, it's good to show highway=cycleway+foot=yes on a specialized foot map.) > So what are the advantages of the change? One scheme that covers the > corner cases along with the most common occurences. And the > disadvantages? Confusion for many contributors, every data user > needing to understand two sets of tagging styles, the most common What are these two sets of tagging styles? I only see one new value, not a whole new style. > cases (the 509k) needing twice as many tags as before, and the corner > cases are still fairly corner needing a small handful of tags. So in Incorrect. You neglected to account for the existing tags on those 509k/425k. There's actually a net gain (reduction) in the number of tags needed. The simplest cases (cycleway/footway/bridleway) are identical, obviously. But now a specialty route which is not a c/f/b is both more intuitive (no highway=footway+foot=no needed) and requires fewer tags (highway=path + snowmobile=designated instead of highway=footway + foot=no + snowmobile=designated for example). At worst, there is only one additional tag needed, not "twice as many", and that's only if you choose to use highway=path for something which is also a c/f/b. And of course, non-designated paths (2nd-4th, 6th and 9th on http://wiki.openstreetmap.org/index.php/Tag:highway%3Dpath/Examples for example) are now only one tag instead of two, as well as being much more intuitive. > proposed upheaval and/or dual tagging regimes is overkill. A way to > tag the corner cases that don't fit in well would have been much > preferable. Upheaval? Dual-tagging regime? Neither have happened here: c/f/b are still perfectly viable, and highway=path simply adds a generic path alongside the existing specific-use paths. And IMO if someone knew of a less disruptive, more intuitive change to make, they should have mentioned it during the 6 months that the proposal was in the wiki. Not that I think anything could describe a non-specific-use path in a less disruptive way than highway=path. -Alex Mauer "hawke" _______________________________________________ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk