The area near the stream is well-walked earth with little grass. There is a certain section which is trampled well enough to consider it an ‘Informal Path’. When adding this path in the iD editor, I get a warning about a disconnected path.
All roads, paths, and ferry routes should connect to form a single routing network. This feature (or small collection of features) is either disconnected from the routing network or only connected to it via a feature that few applications consider routable.
This path is indeed disconnected from any nearby roads or paths, though it can be accessed by foot from nearby roads Donaldson Drive and George Street. Should this path remain disconnected? What are the side-effects of disconnected roads apart from them ‘not being routable’?
The only real consequence is that you will (at least with the current state of routing engines) not be able to automatically generate a route over that stretch of path, which is unlikely to be something useful to start with.
tl;dr IMHO it is totally fine and you should ignore the warning (perhaps add a note tag or similar to it explaining why it is disconnected).
I agree to that in general. On the other side the question is if it really make sense to map such an occasional path stump which may dissappear as fast as the grass is able to recover. I tend to map such fragments only if they are really of permanent nature.
From your description, it should. If there is no “obvious” way to get there, but the path itself is obvious enough to be considered informal, then it is disconnected. Doesn’t mean it’s inaccessible.
Nothing, really. It will probably be ignored, though.
I think it’s still useful for manual routing, to know that there is something visible.
It could confuse a router, if you start from there.
Many years ago I was in Reykjavik and was not able to get a foot route to anywhere. It turned out that someone had tagged a sidewalk around a block and not connected it to anything. So the router (OsmAnd) knew that I was on this footway, but could not find a route to anything.
Often the warning make more sense.
For example in wet areas (meadows, farmland with drains, etc) where aerials show bridges over a stream or ditch and then only those bridges are mapped. A user might then want to get to a bridge to cross a bigger dith to get to a field on the other side. But how does he get to the bridge? A straight line probably is no good because of other obstacles, wet areas etc. Then it is useful to tag some paths to the bridge that would work.
I think it’s fine to leave this type of path disconnected. Validator warnings are just that - prompts that something might be amiss, not that it’s necessarily wrong.
If you did for some reason really need to connect this to a routable network, a link (highway=footway + footway=link) could be added.
I’d say routers should make a point-to-point path segment from the disconnected path to the next closest path, then route from there as normal.
Most routers already do something similar at the end of the journey I think? As in, they don’t fail to create a route if the destination isn’t directly on a path. Rather, they add a virtual path to the destination from the closest known path.
Organic Maps does something quite similar by the looks of it. It ignores the disconnected path entirely, makes a virtual path from the starting point to the nearest connected path, then goes from there:
I have a slightly different take here… while, yes, I agree it’s just a warning and can be ignored, isn’t the whole point of mapping it to help other people find it to access it? Why not just connect it at a reasonable place where you may know it joins to another footpath? There is a break in the trees visible on aerial and there’s another footpath not far down the field where it seems like it could make just as much sense to connect it and resolve the warning. My 2c.
Yes, in part at least, and that is what is mapped. The map requires interpretation anyway, so if there is a disconnected path, I know I have to find my own way there (if I want to use it). That’s also why…
doesn’t really work from the given description. There seems to be no such path (otherwise it could be considered informal and mapped as such). Drawing a connection there would be misleading, as there is no obvious path. See also Tag:informal=yes - OpenStreetMap Wiki.
ah yes, silly me. must be the infamous “trebuchet landings” that launched the OP to this obscure location =D very believable. no possible way to access given the cul-de-sac clearly visible in the graphic. OP states “though it can be accessed by foot from nearby roads Donaldson Drive and George Street.”
like i said i have a different take, and this post has a solution already but, to me, “well-walked earth” is a footpath worth mapping for others to know how to get there in the first place. there’s def a place to connect it, but if it’s too much bother, then ignore the warning and leave it dc and let routers virtually direct looptiloops around and people will figure it out like they have for thousands of years before this platform existed. just, not really the point of this platform, imo isnt it…
No need for ridicule. The informal path can obviously be accessed, noone disputed that, but it is a question of Verifiability - OpenStreetMap Wiki of what to map there. If you were standing there, would you be able to verify that the path as mapped exists?
Walking over a meadow might be physically possible, but that doesn’t necessarily create a path to be mapped in OSM. The purpose of this platform is to map what exists, not map for any specific application.
That’s not always a good idea.Have a look at this route from “Wakefield Kirkgate Station” to “Wakefield Westgate Station”. For illustrative purposes I’m showing that on osm.org, but I actually first saw the error elsewhere (on a Garmin handheld).
The problem is that the “nearest road” to the station node in OSM is actually underneath it. The suggested route requires either tunnelling vertically for about 2m or climbing over a 2m high wall and some security fencing. I’d actually prefer to be told “I could not find a route” rather than given a guess that was actually wrong.
Arguably, it would make sense for the node for Wakefield Westgate station to be instead an area covering the whole station (routing to any entrance would then obviously work) or be a little nearer to the logical centre of the station (near the station building and overbridge). The current layout in OSM seems to be a deliberate decision - there might be a good reason for it, but I’ve no idea what that is. I made the router work with OSM data as is instead.
I first saw that with this router Google Maps – Crossing that terrain will take longer as walking around. The local Mountain Rescue recounts a number of incidents there on their blog. Same route OpenStreetMap or GraphHopper Maps | Route Planner better though do not just use the most complete data but also the right profile GraphHopper Maps | Route Planner
The one path that I would say exists for sure, is the one I’m trying to map. In terms of mapping what exists, and not for a specific application I think it would make sense to leave the path disconnected.
highway=footway + footway=link was suggested by quite a few people here. From what I could gather from the wiki, footway=link is used to connect highways that would otherwise abruptly end.
What is that supposed to mean? It appears to be unsupported, and all the occurrences are in a fairly small area. Do you know anything about the history of why (a few) people started using it?
