openstreetmap
stop automatically adding railway crossing tags when adding tram-road/path intersection nodes
Addresses concerns raised in https://wiki.openstreetmap.org/wiki/ID/Controversial_Decisions#Tram_crossings or https://wiki.openstreetmap.org/wiki/Talk:Tag:railway%3Dtram_level_crossing.
This essentially reverts #7902: Instead of using the undiscussed tags railway=tram_level_crossing and railway=tram_crossing, the regular values of railway crossing tags should be used at intersections of tram lines with roads or paths.
See https://github.com/openstreetmap/id-tagging-schema/pull/596 for a related change to the tagging schema repository.
~~Still to be discussed: Should there be a tag upgrade/deprecation rule be implemented to fix these tram_*crossing tags?~~
//edit: as noted below (https://github.com/openstreetmap/iD/pull/9306#issuecomment-1258412778), automatically adding any railway crossing tags is not good practice. I've changed the PR to not add any additional crossing tags in these situations.
This essentially reverts https://github.com/openstreetmap/iD/issues/7902: Instead of using the undiscussed tags railway=tram_level_crossing and railway=tram_crossing, the regular values of railway crossing tags should be used at intersections of tram lines with roads or paths.
That is not solving main problem: iD asking people to mark fake crossing where none actually exists, see https://github.com/openstreetmap/iD/issues/5990#issuecomment-467876574
As long as this rule is present iD mappers will add fake data. And it is preferable to keep it contained within own tag. (obviously, better would be to remove this rule like JOSM is doing this and rely on manual mapping)
As large part of this tram crossings is misplaced or not showing actual crossing points, this tags should not be semiautomatically upgraded. It is better to keep this polluted data contained within own tag. So this change actually makes things worse.
thanks for the additional information @matkoniecz. I agree that it is best to stop to automatically add any railway tags in these situations, as tram-road crossings are so diverse. Mappers need to choose whether in a particular situation it's best to use a railway crossing tag or no tag at all. I've changed the PR accordingly.
Mappers need to choose whether in a particular situation it's best to use a railway crossing tag or no tag at all.
In that case, this PR still overcorrects by reducing the likelihood that railway=level_crossing or railway=crossing will be added by the mapper when necessary. Adding the intersection node will satisfy the warning, leading the mapper to think there’s no need for a second step to add this tag. This isn’t a theoretical issue: some routers issue warnings to users to look both ways when crossing, but it requires a crossing tag, so they may have a false sense of security when traversing an undertagged crossing.
By analogy, when a roadway crosses a waterway, we can’t be certain that there’s a ford, so the crossing_ways validation rule gives the user a choice of adding a ford or adding a bridge. Similarly, if we’re sure there should be a node but unsure if it should be tagged railway=crossing, wouldn’t it be more appropriate to warn but offer a choice between railway=crossing and an untagged intersection node, similar to my suggestion in https://github.com/openstreetmap/iD/issues/8463#issuecomment-1066069238?
If https://github.com/openstreetmap/iD/issues/5990#issuecomment-467876574 is illustrative of the issue around trams, then it’s worth noting that street-running is not exclusive to trams or streetcars, so for consistency, this rule would need to exempt railway=rail from railway=level_crossing and railway=crossing too. Perhaps it would be less surprising to exempt ways tagged embedded=yes, but you’ll probably notice that it’s merely “in use” – out of necessity rather than due to any editor suggestion.
This situation appears to be an unintended consequence of a 2019 decision to map non-physically-separated railways in finer detail than non-physically-separated footways and cycleways. Maybe the real lesson here is that there needs to be a discussion about digitizing all three types of ways more consistently.
street-running is not exclusive to trams or streetcars, so for consistency, this rule would need to exempt
railway=railfromrailway=level_crossingandrailway=crossingtoo
Also, for completeness, railway=light_rail, which occupies a poorly defined space between rail and tram. Very many light rail lines are street-running in some parts but not others. As I understand it, one of the original motivations for railway=tram_crossing and railway=tram_level_crossing was that the right-of-way rules for streetcars are very different than for heavy rail. But light rail is a mix of the two on a case-by-case basis. It’s similar to how a footway and cycleway can cross each other at either a pedestrian crossing or a cycle crossing, depending on the situation. It’s good to give the mapper a choice, but then we should put the choice in front of the mapper instead of making it easy for them to get yelled at by other mappers.
an unintended consequence of a 2019 decision to map
I can't speak for others, but the Nottingham tram tracks have been fully mapped as separate ways since at least 2013, and I think from around 2011, and possibly earlier than that. It was such a long time ago I can no longer remember the exact reasons for this, but I do remember that there were several of them.
I was unaware of these tags and have reviewed them for the Nottingham tram network. tram_crossing looks not to cause any problems at all, but tram_level_crossing does cause issues where there is street running. In nearly all cases (I may have overlooked some) where the tram is not street running and all significant junctions with street running the junction is protected by tram-activated traffic signals (I'm afraid not always mapped, I'm working on it). Having additional routing penalties imposed on these may not be significant, but for minor road junctions and particularly service roads and driveways connecting to the road with the tramlines. In all these cases, drivers (& cyclists etc) would be expected to observe normal behaviour to check for traffic in either direction and these are no different from roads where there is no tram.
As you say right-of-way rules for streetcars can be vary variable. In many jurisdictions (useful reddit thread here) cars must not pass on the nearside of a tram picking up/discharging passengers (@matkoniecz can correct me, but I think stops on Lubicz in Krakow are like this).
I suspect the best behaviour is to suggest connecting the two ways, but nothing else. Other suggested approaches for handling street running don't seem to be in use.
I suspect the best behaviour is to suggest connecting the two ways, but nothing else.
Why not suggest both – let the user decide?
ars must not pass on the nearside of a tram picking up/discharging passengers (@matkoniecz can correct me, but I think stops on Lubicz in Krakow are like this).
well, obviously - are there places with trams where situation is different? Even Poland when its traffic laws were really pedestrian-hostile was without such mistake.
Why not suggest both – let the user decide?
Because I'm just not convinced there's anything special about these intersections. They have no special signage (as a railway=level_crossing does), are no different from other intersections where we don't have special tags), and on a typical road with street-running we'd typically be missing plenty of other potential examples (every driveway, service road etc). This based on UK, FR, AT, CH, DE and PL tramways. I'm going to trawl though old photos, but foolishly I didn't get any on New Years Day when I was riding the tram.
Also I think there's a strong tendency for users to click yes to suggestions.
Of course if they don't exist they can be ignored, regarded as tagging syntactic sugar on a tram/highway intersection. I don't know if there's any mileage in considering usage of the embedded_rails key instead (not its original purpose I know, but not unadjacent)