I wonder what is the proper fix. Remove ford=yes
? Remove note connecting stream and road?
Finding all ford=yes
edited by this user is fairly easy, the problem is that I am not sure what to do with them.
I wonder what is the proper fix. Remove ford=yes
? Remove note connecting stream and road?
Finding all ford=yes
edited by this user is fairly easy, the problem is that I am not sure what to do with them.
Most of them are culverts, as I wrote in fourth message in this thread MapRoulette Critique - #4 by Jarek. In some cases the imported waterway is just wrong and should be removed instead (but care should be taken to use several imageries as many of them are more of a ditch that only has water in it in spring or after heavy rains).
Again, this is something that’s obvious to local editors, something that’s probably less clear to a mapper dropped in a region they don’t know by a tool that encourages them to “solve warnings” in other tools.
As you open the discussion, Aside from any problem with MapRoulette, we could simplify the way to represent culverts for minor infrastructures.
Representation of a culvert could be simplified by adding a simple node where a highway intersects a waterway as this structure is related to both. Would node (highway=culvert) could do it ?
There are roads where you have hundred a minor culverts infrastructures for each stream and a lot of warning when the culvert is not documented (example way > waterway=stream, tunnel=culvert, layer=-1). Dont you think that this overTagging tendancy for some infrastructures is contributing to worsen the problem ?
Representation of a culvert could be simplified by adding a simple node where a highway intersects a waterway as this structure is related to both. Would node (highway=culvert) could do it ?
I don’t think that’s a good idea. They do not intersect hence there should be no common node of a highway and a culvert running under it
They do not intersect, so do not connect a waterway to a road unless it’s a ford, but you can indicate on a road that at that place a culvert is underneath. I wouldnt use the word intersect for that, though.
not sure is it any simpler in practice than current mapping
Well, a highway and a building do not intersect either, but we map them with a shared node, see Tag:tunnel=building_passage - OpenStreetMap Wiki.
I think a shared node would be ok, as a first step before refining to a way (without a shared node then) because often you dont know how long a culvert is (where it starts and ends), so splitting the waterway to create a culvert would be guessing. I see the culvert more like a waterway feature than a highway feature though.
(I asked to split this discussion into a separate thread)
Well, if the waterway is drawn underneath the road anyway, I see no benefit in putting the tags on an extra node on this waterway.
personally for me in case of File:Wullersdorf Pfarrhof Durchfahrt.jpg - OpenStreetMap Wiki building in fact is intersected by road
contrast to case of a carriageway where culvert is below it and is not surrounded by it on its sides
disclaimer: I am not a native speaker
There was a waterway, somebody added a road over it with a small culvert under it to let water pass through. Then Climate change did come with more and more sudden surge of water and the culvert and road were washed. A waterway or road problem ?? But for mapping roads, some contributors decided that they did not want to take care of this problem. Those that map waterways dont bother either …
Most of our “culverts” in Nederland are just overflow pipes from one ditch to another. If a farmer wants an extra access point to their land they just dump some earth in the ditch, then throw 5m leftover pipe 15cm in diameter, then fill up to passable level and put a swing gate on. I have a hard time tagging that piece of pipe as waterway and tunnel.
On the other side we have navigable culverts as large as bridges (just with concrete bottom) where I find current culvert tagging (as a way) utterly inadequate.
End rant. Sorry.
building=
is a physical structure. Both highway=
and waterway=
are routable functional topologies. Not to mention the area vs line difference.
This concept was already discussed before, ignoring naming of =simple_brunnel
Why, you can at least draw an area man_made=tunnel
+ tunnel=culvert
Then I suppose that we will find contributors motivated to do it and never talk of the problem again …
In this picture alone, authoritative data indicates 5 points where roads pass over a stream or a ditch.
True. Did that a few times. Also added the water as an area, because it usually connects water areas, sometimes also containing the waterway if it’s needed for naming or routing. Thinking: the culvert/tunnel is the concrete structure, the water flows in it.
For the pipe version, I would prefer not having to tag a waterway=ditch/stream/drain for a connecting pipe, even if it represents flow continuity of a named waterway.
However, I don’t think these ideas of mine will get much traction. Most mappers think it’s utterly pointless to map culverts or small waterway connections.
they do intersect