|
What is the best way to handled co-aligned trails (i.e., when two trail systems share a common path). Is it to defined multiple Relationships for the path, one for each trail systems that are co-aligned over this path? So for example the trail segment (Path) with two co-aligned trail systems would have Reationship-1 "Hiking Path = Highlands Trail" and Reationship-2 "Hiking Path = Lake Hopatcong Trail". Thanks! --Rob
showing 5 of 12
show 7 more comments
|
|
Yes, multiple relations are the way to go. |
|
The physical path you walk on should only be mapped once. This can be as a single way or multiple ways drawn end to end. These ways should not overlap unless there is e.g. a bridge with a separate walkway that goes along underneath it. If multiple routes use that same path they can each be represented by separate route relations referencing the same osm ways where necessary. Route related tags should not normally be present on the ways that make up the route themselves. "Route related tags should not normally be present on the ways that make up the route themselves." What a strange piece of advice, usually route related tags go on the route relation, not the ways.
(25 Apr '21, 09:14)
yvecai
1
@yvecai It's written as a negative, but "Route related tags should not normally be present on the ways that make up the route themselves." and "route related tags go on the route relation, not the ways." mean the same thing?
(25 Apr '21, 11:11)
SomeoneElse ♦
I am really not clear on this comment relative to what I have proposed. I proposed defining multiple relationships of the Path (singular) with Hiking Routes (plural). Each co-aligned trail system belongs to a unique Hiking Route.
(28 Apr '21, 11:46)
Rob009
|
|
Ah, sorry, I missed the negation even if I copy-paste the quote! [shame] |
|
Other advice I received (outside of this thread) was to set the Role of the Hiking Route as "alternative" to indicate the secondary nature of the coincident paths. The through trail is assigned Role=Main and all others have a Role=Alternative. This could be correct depending on the how the trails are set up. To nitpick on wording though: roles apply to elements of a relation and exist within the relation definition, they do not apply to the whole relation.
(28 Apr '21, 14:10)
InsertUser
|
For completeness, the relevant bit of the "Highlands Trail" appears to be https://www.openstreetmap.org/relation/6335422#map=13/40.9476/-74.6900 (I suspect that the name isn't really "Highlands Trail (Teal Diamond)") and the relevant bit of "Lake Hopatcong Trail" is https://www.openstreetmap.org/relation/12621105#map=14/40.9433/-74.6783 . Both are visible at https://hiking.waymarkedtrails.org/#?map=14!40.9425!-74.681 . I suspect that the names on https://www.openstreetmap.org/way/613120946 and https://www.openstreetmap.org/way/277931899 shouldn't really be there.
Yes, the Hiking Route name is "Highlands Trail (Teal Diamond)". I just shorted it for the purpose of the example. Also, this particular Path was already on OSM identified as "Unnamed" (Name = "Null"). I named the path "Highlands Trail" (it already had relationships with the "Highlands Trail (Teal Diamond)" and the "NJ Long Trail" Hiking Routes).
I added "Brooklyn Mt. Tr. - Lake Hopatcong Trail spur", which in fact another trail system (Route) that is in the area. Why do you say "it shouldn't really be there". This is a Spur of the Lake Hopatcong Trail system, not the Highlands Trail (I volunteer for both). The addition of the complete Lake Hopatcong trail system is pending (the spur you mention was already on OSM identified as "Unnamed" (Name = "Null")).
Currently, https://www.openstreetmap.org/way/613120946 is one very long piece of trail. I suspect that there are bridges yet to be be mapped, and changes of surface, and changes of width, etc. Together, all of these form part of the "Highlands Trail", and that's mapped as https://www.openstreetmap.org/relation/6335422 . Each individual way doesn't need that name as well - quite possibly they will have other, different names (a bridge might have a name "Foo bridge" for example, and https://www.openstreetmap.org/way/766379385 already has another name "Pine Grove Road").
"Each individual way doesn't need that name as well". Why do you assert this? Without naming the individual Paths that constitute the Route you end of with "Unnamed" trail appearing for these trail segments in applications like Gaia GPS, which use OSM for their base data. (in fact this is the very issue that led me to making OSM updates - unnamed trails and missing trail segments in Gaia GPS)
If applications like Gaia GPS cannot handle names of routes then they should change their way of data processing. There are so many applications out there using OSM data. We cannot possibly try mapping for all of them.
There are many reasons why a trail name should not be mapped on an individual way. As you noted yourself often several trails with different names use the same path. How would you name that path then? Also often segments do have individual names that differ from the trail's name. As an extreme example take a sightseeing route in a city following the road network. Do you want to put "Great City Walk" on all the roads that are rightfully already called "Main Street", "Washington Avenue" and "Old City Square"?
The whole idea of using a route relation is to do this abstraction between the physical way and the route following it.
OK, I take your point, but what exactly is that harm of populating the "Name" attribute of the Path? I sense this argument is an example of "the best is the enemy of the good". In other words, what harm is there of naming a trail segment to correspond to a common name of the trail in the area? Yes co-alignment complicates this but as long as there is agreement on the path naming syntax between the Hiking Route operators, do it really matter?
BTW, I will reach out to Gaia GPS regarding the apparent lack of use the the Hiking Route relationship. -- Thanks!
Yes, it does matter. The "name" of a way in OSM is the name of that individual way, not the name of any relations that it happens to be part of (and as you've said already, ways can be part of multiple relations). If people put all sorts of random guff in the name field (which probably includes " (Teal Diamond)" too) then how will people know what the actual name is?
To be clear, I am not arguing against the use of the Hiking Route to define the trail system. What I am arguing against is your apparent dismissal of the Path name. Yes it is a freeform field and that means no standardization, but so what if this uncertainty is managed by the local Operators.
I have asked Gaia GPS about there use (or lack there of) of the Hiking Route relationship of individual Paths
I advise against generally applying route names to individual ways. Nevertheless, there may be occasions where paths are actually named like the route or at least referred to by that name by the locals. By my experience this is sometimes the case for short sections of trails passing through a village.
In that case it is ok to name the way directly. As so often in OSM it is a subjective judgment by the mapper what is appropriate.
OK, thanks I will take this as a "best practice" not to blindly name each way with the name of the Hiking Route (or derivative there of). As previously noted, I have escalated with Gaia GPS their apparent dependence on way naming.
The relevant bit of the OSM wiki is https://wiki.openstreetmap.org/wiki/Names#Name_is_the_name_only .
Thanks for the Wiki pointer. Ok, here is a very specific question. The trail system (Hiking Route = Lake Hopatcong Trail) was several shorter feeder trails called "spurs". These Lake Hopatcong Trail spurs all have unique names, like the "Leo Avenue Spur" (meaning it services hikers entering the trail system from the street Leo Avenue). What is the must appropriate approach to associate "Leo Avenue Spur" with its way (a way that has a relationship with Hiking Route = Lake Hopatcong Trail). Is it to use the Name tag in this case, as "Leo Avenue Spur" is specific to this item?