Proposal talk:Hiking trail relation roles/Archive 1

From OpenStreetMap Wiki
Jump to navigation Jump to search

"All hiking trail relations should be updated. "

Why? allows leaving role untagged and many trails have only main section, without any of special sections Mateusz Konieczny (talk) 10:35, 6 December 2019 (UTC)

Thank you, I changed it. --Mfbehrens99 (talk) 11:04, 6 December 2019 (UTC)


Mateusz Konieczny (talk) 11:14, 6 December 2019 (UTC)

« A only officially marked trails should be added to the relations! »

OSM is about mapping the real word not the official world. So I’m againts the "officially". And against "marked" too, a trail not marked can be added to OSM as long as it exists. (Florimondable)

Unmarked trails should not be added to OSM (except rare extreme cases). OSM is not a place to save things like unsigned "my favorite trail" or "good running route". Mateusz Konieczny (talk) 11:13, 6 December 2019 (UTC)
Maybe "unofficial signed trail should not be added to a relation of an official trail, but rather mapped separately" would solve part of the problem? I agree that unofficial signed trails can be mapped Mateusz Konieczny (talk) 11:18, 6 December 2019 (UTC)

Actually it is not about any officially marked trails but more about marked trails. So when you decide to mark your own way, you should map it aswell. The main point is that you should only add ways into the relation that are still marked. So if there is a viewpoint that can be reached on a path without markings than they should not be added into the relation. --Mfbehrens99 (talk) 11:21, 6 December 2019 (UTC)

Ok, the point is : Only add to the relation secondary trails that are realy part of the trail route, not made up or other trail routes. Please rephrase this way, because it was not very clear for me. Florimondable (talk) 12:45, 6 December 2019 (UTC)

I changed it. --Mfbehrens99 (talk) 13:01, 6 December 2019 (UTC)



I think nodes should be a separate proposal. There will be enough to do with the above without introducing yet more complications. I will list some of these below. Warin61 (talk) 08:41, 7 December 2019 (UTC)

+1. For example, for me "overnight" role seems to be a horrible mistake and enough to turn me against the entire proposal. Roles for ways seemed OK to me Mateusz Konieczny (talk) 08:49, 7 December 2019 (UTC)

I removed it from the proposal. You are right that this is probably too much. --Mfbehrens99 (talk) 10:53, 7 December 2019 (UTC)

overnight protest

Please no. What is the benefit of putting all accommodations near trail into a hiking route relation? What is the benefit of making every accommodation and alpine hut in Tatra mountains member of 30+ hiking routes? See

It seems to be case of something easy to automate (check whatever tourism accommodation is within X minutes from trail) and asking mappers to tag it manually in way that will make editing far harder without a good reason. Mateusz Konieczny (talk) 08:48, 7 December 2019 (UTC)


On a two way route there will be two 'starts', that may be confusing. Is it an error, or is it normal? Warin61 (talk) 09:01, 7 December 2019 (UTC)


An 'end' can be the start or finish - it is just one 'end' of the route. Complication - where a single route can be 'started' at any convenient point - so there may be more than 2 'ends'. Warin61 (talk) 09:01, 7 December 2019 (UTC)

Already existing features - parking/toilets/camping/hotels etc

Why make more work when these things already exist? The may also be mapped as enclosed areas, so possibly not only nodes. Why not simply include them with their existing values as the role in the relation? This would mean that any existing or future feature can easily be added rather have have to think of something, such as boot cleaning stations. Using the existing value is less work and less error prone. Warin61 (talk) 09:06, 7 December 2019 (UTC)

Note I am not at this time supporting the idea that existing things such as car parks need to go into these route relations. ~!Warin61 (talk)

Consequences for export and other uses

It seems to me that use of these roles leads to relations containing non-contiguous trails. I would call those relations collections rather than routes. Processing non-contiguous routes presents extra challenges for processing such as exporting routes and making elevation profiles.--Peter Elderson (talk) 10:59, 6 December 2019 (UTC)

The problem today is that often routes are already collections of all the ways that make up such a trail. In future this would be easier because then you can only take the ways with the role None or main and generate a elevation profile from this contiguous main trail. --Mfbehrens99 (talk) 11:09, 6 December 2019 (UTC)

Exactly. So if you enter alternatives with roles into the main route relation, you exclude the alternatives from export and profiles. Correct? --Peter Elderson (talk) 12:59, 6 December 2019 (UTC)
That was the plan. Would it change anything when there was a big relation holding small subrealtions where the roles are assigned on? --Mfbehrens99 (talk) 13:44, 6 December 2019 (UTC)
The roles do not lead to non-contiguous trails. They only document sections where the route is non-linear. Non-linear routes can be found quite often in reality. Where a route is marked it should be mapped as such. If that means to add an alternative, so be it. Artificially splitting routes at deviations does not help to improve the data quality. Also, the processing point is mute. Adding roles significantly helps with processing. Lonvia (talk) 22:54, 6 December 2019 (UTC)
For export, e.g. a GPS file, it could be that the main route is exported as one file and the alternate, excursions etc. exported as other files .. this then transfers all the data in a reasonable form. Other uses? Probably a similar approach. Warin61 (talk) 23:44, 6 December 2019 (UTC)

"alternate" or "alternative"

As Lonvia pointed out: "alternative" and "alternate" have almost exactly the same number of uses. We could have the discussion if one of the two should be deprecated but I tend to think: why bother. (Thread: Cycling too?)

I would write that maybe one is the right thing that should be used when you are tagging new routes and the other can be accepted by some applications in the beginning of these new roles. I don't think it is a good idea to have two "different" roles for exactly the same thing. I would take alternative as the new correct role and use alternate as an accepted role for some years until there was enough time to change it. --Mfbehrens99 (talk) 06:11, 7 December 2019 (UTC)

I think that it is a good idea but note that at least some people will be against it just because it deprecates something. Mateusz Konieczny (talk) 08:51, 7 December 2019 (UTC)
The change is infinitesimal, even I could not object. :) Warin61 (talk) 09:36, 7 December 2019 (UTC)
About which change are you talking?--Mfbehrens99 (talk) 10:51, 7 December 2019 (UTC)
A change from "alternate" to "alternative", or from "alternative" to "alternate". To me the difference is so small that either could be used with no loss of comprehension. Looks like "alternative" wins for the precise meaning. Warin61 (talk) 21:23, 7 December 2019 (UTC)
I'm assuming this may be a difference between usage in en-gb and en-us. I would naturally refer to these as alternative routes. US English tends to provide a broader range of meanings to alternate than en-gb (e.g., deputy and as a synonym of alternative). In context I think both are clear, and whereas I have a personal preference for "alternative", I have no problem understanding "alternate" here. By all means recommend one on the wiki page, which may be followed up by editor maintainers, but I really don't think it's a huge issue for data consumers. SK53 (talk) 14:59, 7 December 2019 (UTC)
In the UK, alternate would mean that the direction of the route changes. As tags (and likely also relation roles) should be in British English and because alternative is also understood in the USA, i'd only choose alternative. --SelfishSeahorse (talk) 15:33, 7 December 2019 (UTC)
I like alternative also better. With the support of SelfishSeahorse's argument I will decide that alternative is now the correct role name. --Mfbehrens99 (talk) 16:13, 7 December 2019 (UTC)

Is it OK to tag entire trail as approach/excursion/alternative?

is black trail supposed to be tagged as excursion in its entirety?

Some trails can be in entirety described as approach/excursion/alternative. Is such role supposed to be used on them? In Poland trails are typically split into multiple rather than having trail cross with itself. See for example Mateusz Konieczny (talk) 10:31, 6 December 2019 (UTC)

Yes, all the ways making up the path to the lake should be tagged as a excursion. --Mfbehrens99 (talk) 11:03, 6 December 2019 (UTC)

I would put the ways in a separate (sub)relation and assign the role to the subrelation. Is that within the scope of this proposal?--Peter Elderson (talk) 12:52, 6 December 2019 (UTC)
I would appreciate if this proposal would not be hijacked by the subrelation discussion. This proposal has nothing to do with that. It is about marking the function of the section of a way. Lonvia (talk) 09:13, 7 December 2019 (UTC)
Pelderson, why subrelations would be applicable at all in this case? In this case black hiking route is not a subrelation of any other route Mateusz Konieczny (talk) 13:26, 7 December 2019 (UTC)

Mateusz, regarding your example, this looks like a completely independent route to me because it is marked in a completely different style. We are talking here more about excursions that use the marking of the main route. Lonvia (talk) 09:13, 7 December 2019 (UTC)

It seems that such case should be explicitly mentioned in the proposal. Note "Yes, all the ways making up the path to the lake should be tagged as a excursion. --Mfbehrens99" alternative view. For me both are OK, I would just want to have it decided as part of a proposal. Mateusz Konieczny (talk) 13:27, 7 December 2019 (UTC)
I add you sentence to the proposal. Can this thread be considered as solved? --Mfbehrens99 (talk) 13:57, 7 December 2019 (UTC)
To confirm: So separate hiking route, with its own signage (black one in this case) would not qualify for excursion role? So black one here will have ways with a blank role? Mateusz Konieczny (talk) 18:47, 7 December 2019 (UTC)
The black ways should not even be added in the relation because it has not the same signage and is therefor not belonging to the trail.
Only add secondary trails to the relation that are realy part of the trail route, not made up or other trail routes. --Mfbehrens99 (talk) 07:41, 8 December 2019 (UTC)
Resolved: Mateusz Konieczny (talk) 20:45, 8 December 2019 (UTC)

alternate vs alternative

"alternative (alternate is also accepted)" seems to suggest that "alternative" is preferred. But row later "alternate" form is sole available in "alternate:forward and alternate:backward" Mateusz Konieczny (talk) 20:51, 8 December 2019 (UTC)

Resolved: Mfbehrens99 (talk) 21:16, 8 December 2019 (UTC)


For nodes (like e. g. guideposts) I would like somebody to propose some role names. They are actually also part of this proposal.--Mfbehrens99 (talk) 13:40, 6 December 2019 (UTC) --Mfbehrens99 (talk) 13:55, 6 December 2019 (UTC)
Nodes should be a separate proposal. Warin61 (talk) 21:09, 15 December 2019 (UTC)
True --Mfbehrens99 (talk) 08:19, 17 December 2019 (UTC)