I focus on getting all the signposted hiking routes in my area into OSM. Every now and then in city’s they will cross through a pedestrianized squares like here in Rapperwill https://www.openstreetmap.org/way/26262225.
Now I am considering how to map this in a way that it is clear.
Adding the whole area to the relation would not be correct, since only part of the square will be used, but I see that is being applied to other routes.
I also searched within the forum, but was so far unsuccessful finding a topic related to this.
I am considering if ways without a highway tag (thus invisible) would be an option. In this way the route relation can be drawn without a actual highway, if possible. I would be pleased with your considerations :D.
Is it a legal restriction that only part of the square will be used? If not, including the whole square is probably the correct thing to do. It is up to the routing software to realise that the straight line across it is the best route to follow.
The chances are that any source that is more specific than that is subject to copyrights.
I don’t have an answer but I have been having the same issue and would be very interested in solutions. Often it is obvious on the ground how the area should be crossed to follow the trail, but there is no specific physical “footpath” within the area.
Other examples I have come across:
Part of a trail goes along a beach. Typically it is obvious on the ground that the trail follows the edge of the beach, there may even be waymarks along the edge to indicate this, but not a visible footpath. So adding a path doesn’t seem right as there is nothing visible on the ground. But adding an entire beach when the trail just follows a narrow strip doesn’t seem right either.
Part of a trail follows a river bed that is dry most of the time.
The trail crosses a car park. The car park may be a “rough and ready” area of ground so there are no visible parking aisles or footpaths to map. Again it is often obvious that the trail follows the edge of the car park, and adding the entire car park to the route doesn’t seem right.
The trail crosses a grassy field. There may at times be a faint path visible but it is likely to disappear as grass grows.
That’s an interesting idea. But would the way have to be without any tags? I think this would be seen as an error by Q+A tools and there might be a high risk that somebody else would delete it.
I wondered about using highway=path (or footway) with trail_visibility=no, to indicate that there is an implied path to follow but it is not directly visible. But I think trail_visibility=no is more often used for difficult trails where navigational skills are required, which is not what we are talking about here.
I am not sure if sources are really relevant here. The general issue is just as likely to arise where a mapper has surveyed the route in person. That is certainly my situation - I know exactly where the route is because I have followed it, I just don’t know how I should map it.
Where a route crosses a square I would add the square to the relation as long as it is mapped as a highway=*. That is probably not the ideal solution but with current mapping conventios the best in my opinion.
If these are all areas that do not carry any highway=* tag a new way should be added where the trail likely is. Only highways are surely considered wakable ways by all data consumers. You can add trail_visibility=* to indicate how poorly visible a trail is.