As a hiking enthusiast I would like to improve OSM for hiking/the outdoors.
As many long distance trails are still missing on OSM, I was thinking to start a community to do this work together with other mappers.
So I was wondering if there are other enthusiast here and what you think of the following approach:
Identify missing long distance trails worldwide (iwn and nwn)
Get GPX with permission to use in OSM
Divide and keep track of the task using HOT Tasking Manager
Have several projects in this way
As far as I found, there is not a similar kind of project right now.
I suspect that this is very much country-dependent. Which countries are you aware of that have lots of missing trails?
I’m not convinced that “just vaguely adding the route” is of much benefit on its own. What is useful is the extra data that OSM tends to provide - gates, stiles, surface, difficulty (e.g. sac_scale) etc. If you could get (or survey yourself) missing trail data then that would be far more interesting.
Count me in! Nederland.
Nederland has hiking project, Belgium, France, and Germany, and in the UK some mappers are active, though I’m not aware of any projects.
Most mappers are mainly concerned with routes in their own country.
I have done some work on iwn routes in Europe; that’s not easy.
Most iwn routes are superroutes consisting of national sections, which may be nwn routes in their own right, often sectioned themselves. The hierarchy can go deep, and there are many pitfalls.
Still, why not see how far we can get!
About gpx: you need to make sure that it’s an authoritative gpx, and that it represents a true signed route. In the end, all routes need to be verified in the field.
In Nederland and Belgium, we have contact with national operators, who keep their downloadable routes updated. We have tools to compare their gpx with the corresponding OSM-route relation, usually we find quite some differences which are then solved and improved at both ends.
I am not familiar with HOT tasking manager.
Nederland and Belgium have their own hiking project pages on osm wiki.
Nederland’s hiking project is mainly an incredibly long table of waymarked hiking routes with a completion indication. Some routes are never updated, though. Nwn and iwn routes (the Dutch sections) are complete and are reasonably well maintained.
True. In Nederland, we have a national operator who have given us explicit permission to use their data (long distance routes an node network) for OSM. Still a lot of hiking route operators have not given overall permission, so we need to ask for permission for each and every route they offer.
If there is no permission, survey is the only option. After that, we can check for differences and contact them about that.
When they realise that the route will appear on OSM anyway, usually permission is granted to use the download or their map, and they profit from our surveys and from the website link we put in the route relation.
That’s good. The opposite end of the scale is of course the crazy French situation, where the FFRP claims copyright over its routes (and is backed up in French law) so you can’t even survey them yourself.
“permission to add bare route data” is a substantially inferior alternative to actually mapping the detail that occurs along the route. You’d maybe do it if you didn’t have a particularly active OSM community locally (proportional to the length of unmapped routes), but mapping hiking routes is among the least import- or armchair-suitable activities in OSM.
Unless, of course, you consider the route itself important information for the hiking world.
When I started to do this, the OSM information for long distance routes in Nederland was in very bad shape, although a bunch of mappers were active. Not complete, and most were hopelessly outdated, and relation integrity was bad. Now, all the routes of the national operator, and many others, are complete and most are up to date, and there is a system of maintenance emerging. Also tooling is being programmed for maintenance of the routes in OSM.
Most of this was armchair mapping to get it complete, then survey to improve quality, and quality projects involving operator maintenance teams (the people who do the signing).
Initially, we entered low quality data, and we knew that. But we knew what and where to improve. The alternative was to keep on entering incomplete high quality data, knowing that say a year after we were done with a route, the route information would decay because of mutations of the physical routes which were not reflected in the OSM route relations.
Same story for the node networks that cover almost entire Nederland. In fact, it’s one node network for the whole country, crossing borders into Germany and Belgium. The complete network is available, because we got permission to use the national database for that, and there is now a universal Node Network planner based on OSM.
So, if you can use good quality route traces from the acrtual source (operator), that’s a good start, after that you’re always going to have to survey for maintenance. In Nederland, all mutations are processed by the national operator on request of the actual field operatives who do the signing, which means they are survey reports. So in fact, we do quality assurance surveys on top of that.
I think adding the trail itself is a very good armchair activity for two reasons. First of all, adding trails and especially the relations ask for more profound knowledge of OSM, which not all hiking organisations have/ have the time for. Secondly, once a trail is added, it is way easier and lowkey to improve it with specific information (also by phone/on the road).
Concerning the GPX files: seems that it would be good to have a place where several files can be added, so that mapping can be based on more than one source. I think for most trails it should be doable to get several GPX files, as many hikers record their hike but are not aware of the value of the data. The GPX should indeed be correctly licensed. And only waymarked trails of course.
Concerning the locations: there might be some parts in the world where most trails have been added, but for most countries this is not the case. And this would also be a global project, so also helping in countries where there are less mappers.
That is one of the pitfalls. Most “innocent” hikers do not exactly follow the waymarked trail. They stray, they lose track, they go off the trail for coffee, they have different approach and end routes. You will need additional assurance that they have followed the trail, or where they have exactly followed the trail. And even then they just forget what they did exactly. You cannot count on a heat map either, you are lucky to get one partial gpx.
In my experience, you need an authoritative source for the official waymarked route, available as open data or with specific permission to use it for OSM (which also means, others can copy and use it, even commercially, specifying OSM as the source, not the actual source).
About that last remark, if permission is granted, the source will fall under “OpenStreetMap contributors” and should be mentioned on the Contributors - OpenStreetMap Wiki page.
Then maybe it should have its own project page, also listed on Mapping projects - OpenStreetMap Wiki
You could link to ongoing projects and community teams from there, and to the HOT project page (or whatever HOT offers).
I would not gather gpx-files. gpx is a transport format, to be discarded after use, because it always has frozen information which becomes outdated very quickly. The same goes for all the routes on all the sharing platforms. Frozen outdated history, fun to have but not a good source. Usable within a project as a temporary store, then better removed in favour of real life data and proper maintenance.
(Ok, ranting doesn’t help either…)
Trying to do “everything before everything else” sounds like a recipe for unforeseen problems occurring all over the world, rather than in an initial pilot area. I would absolutely suggest trying an area local to you first (so that you can verify results locally by survey).
This sounds like another bad idea - if there are few mappers locally there are few people to spot armchair-edited errors.
There have been a couple of examples recently (see for example here and here) of apps misrepresenting data in OSM and/or data in OSM misrepresenting the real world. Having people add things as e.g. “highway=path” with no other tags in potentially dangerous terrain is a spectacularly bad idea. Adding existing “highway” features to routes is far less problematic of course, but still not trouble-free - routes can contain errors (e.g. accidentally routing across private land) and being added to a route relation may misrepresent that.
Routes can also be blocked or move (like here and here) - adding to a route without a survey will result in invalid data being added to OSM.
Generally I would be supportive of any initiative to improve mapping of hiking trails - my following points are intended to be constructive.
I’d need to see more detail on this particular proposal to judge it. My two main doubts would be about the global approach and the “armchair mapping approach”.
On the global approach: is the idea that people would map outside their own countries, or that the project would co-ordinate local communities? My concern would be that communities that haven’t mapped their long distance trails don’t have active members with the enthusiasm and knowledge for something like this. But the issues are likely to be so country-specific that it would be difficult for outside mappers. For example in Spain you would probably need to investigate and contact the “official” sources for each of the 17 autonomous communities separately.
Is your proposal focused on creating missing route relations where the member ways already exist? Or also creating highway= objects that are not on the map? The latter seems difficult as an “armchair mapping” task - footpaths are often much harder to map from aerial imagery than roads/streets, and as others have said, missing details such as whether a path is private or not can cause serious problems for map users.
That depends on how closely the original mapping matches what is mapped on the ground. I have sometimes found it frustrating to fix up route relations created from “official” sources by mappers who clearly never walked the route. Maybe routes added in this way should be systematically tagged with something like “survey:date = never” so that at least other mappers would understand the potential issues.
I would share Peter’s concerns about user-contributed GPX. When I map hiking routes using my own GPX traces, I know where I have diverged from the waymarked trail, but it would be impossible for anyone else to interpret that. Especially as the focus here is on long distance routes, probably mostly multi-day routes, where people are likely to leave the trail to eat and sleep.
Unfortunately the official data sources are scarce or non-existent leading to a situation where we can only map by gpx files from organizations tending to the particular trail (when we get permission or CC0 data), a raster map that we get permission to draw from or on foot following the signs on the ground.
Currently it seems to me that the authorities as well as the main outdoor associations in Sweden are unable or unwilling to help with mapping the hiking paths in a non-proprietary way. This is very different from e.g. Norway where there seems to be very good coverage and high-quality data already in OSM.
@Meinrad_Samuel Do you have enough to continue setting up a global project?
I think the global effort would be to support national/regional teams setting up and managing their operational projects, and connecting to surrounding regions to bring them together as an iwn network.
PS I am reluctant to use the Humanitarian Tasking Manager for just recreation. Feels like devaluation.
What I take as the most important message for now is that we need a pilot project, where paths are already added but not the relations. I will see if I can find a suitable long distance trail that is not on OSM yet which can serve as a test.
Secondly, close collaboration with hikers and mappers in the specific area is, rightfully so, pointed out here. I totally agree, as trail data is only valuable when it is (kept) up to date. The specifics of this project are still very open, so it might just happen that the focus will be on creating more connections between experienced mappers and hiking organisations. Lets see!
If anyone is already interested in forming this project, please sent me a message!
