On Wed, Jan 2, 2019 at 8:13 AM Peter Elderson <[email protected]> wrote: > Sometimes it would, sometimes it would not. If the node actually represents > the start of the trail, it is already in the relation because it is part of > the way that belongs to the route. In the situation that a trailhead node > represents a named cluster of helpful facilities/amenities in the vicinity of > several trails or networks, you wouldn't want to add it to all the relations, > because a. it's not actually part of the routes and b. maintenance of all the > routes would be quite error-prone and not really intuitive. > > A site relation has been suggested for the more complex trailheads. You would > include the node there, the parking(s), the information booth or guide > stands, maybe PT-stops, possibly the route relations you can access from the > site... > > Mapping a trailhead node as I suggested does not stand in the way of more > complex options. My idea: begin with the simplest common element which > supports all the other options.
At the risk of repeating myself: I think I'd need more concrete examples before I'd support such a proposal. I think that we have people in this conversation with different cultural expectations of what a 'trailhead' is. My northeastern-US definition is, "anywhere that you get on and off a trail", so usually there's parking, and perhaps a notice board or a register book to sign, but I don't expect many more amenities than that, and sometimes not even those. It may happen that a trailhead is in a developed facility in a park (such as a ranger station, recreation ground, campground or visitors' center), or even in a populated place, but in that case I think of the amenities as associated with the other facility and not with the trailhead. (Except, of course, for the trail-specific ones such as notice boards, signposts and registers!) If what's under consideration is 'a NAMED place to get on and off a trail,' then I know of only a handful of trailheads anywhere me that have names other than the names of geographic features that they're near. (The "Route 23 trailhead" or the "Roaring Brook trailhead" are typical - they are simply informal descriptions, not real names.) There are a handful of exceptions, like 'Sled Harbor' (near 42.5237 N 74.5629 W) or 'Elk Pen' (https://www.openstreetmap.org/node/1305445030) but they are actually described well by place=locality, since they name the place, not the trailhead. Historically, Elk Pen was where rail tycoon E.H. Harriman kept the elk for his private hunting preserve, and Sled Harbor was where loggers stored their sledges in the summer months. 'Named uninhabited place' is a good description of these. If 'trailhead' degenerates into 'any intersection of a trail and a highway' (which is what it is in that National Park Service database) then it's kind of redundant. It appears to me that the Europeans have a more specific idea of what a 'trailhead' is - but I don't quite understand that idea, and I suspect that's because there are no trailheads of that sort near me, despite the fact that I'm within an hour's drive of hundreds of hiking trails, including a handful of 'big name' long-distance ones. I'm not against the proposal, necessarily, but I'm far from convinced that everyone is reading from the same page, and I'd like to avoid the risk of a false consensus. _______________________________________________ Tagging mailing list [email protected] https://lists.openstreetmap.org/listinfo/tagging
