Talk:Tag:tunnel=culvert
Inferred culverts to avoid editor warnings
In places near my home in South Carolina where waterways are missing, I have added them from the National Hydrology Database. I often find other editors come and "fix" the resulting highway/waterway crossings with culverts, when I do not have enough information to know how the highway and waterway cross and doubt the other editor did either. Today I ran across a freshly mapped culvert with a note "Culvert inferred, unlikely to be real - requires local observation for resolution". This editor is admitting that he has added data to the map that he believes is false, nearly certainly to make some automated debugger happy.
Are there any good practices documented on the wiki regarding resolving waterway/highway interactions in general? If not, should we mention here that features should not be added to the map in absence of information solely resolve a software warning? I suspect a large number of the culverts in the US were in fact mapped based on this sort of inference, since the highways and waterways have largely been populated by TIGER and NHD and many rural areas don't see a whole lot of editing by people with local knowledge.Bdentremont (talk) 20:10, 21 January 2024 (UTC)
- "Are there any good practices documented on the wiki regarding resolving waterway/highway interactions in general?" - not sure, but "Culvert inferred, unlikely to be real - requires local observation for resolution" type of edits is blatantly and clearly wrong. Ideally it would get reverted, with changeset comment and escalation to DWG if that editor continues Mateusz Konieczny (talk) 10:16, 23 January 2024 (UTC)
- I agree. I've commented on that note and am awaiting a response. If I don't get one by this soon, I'll leave a changeset comment suggesting reversion. Unfortunately, I think the underlying behavior is broader than this individual, as tunnel=culvert seems to appear in mass following any occurrence where I've added waterways that cross roads. However--outside of the editor admitting he is likely wrong--its hard to critique a behavior that the project's editing software seems to encourage with its warnings.Bdentremont (talk) 03:14, 24 January 2024 (UTC)
I'd propose something like this for this page and will likely propose something similar for bridge if people think this one is appropriate. Is there a better forum to get opinions from other editors than this talk page?
Waterways crossing highways should not be tagged with tunnel=culvert unless a culvert is known to exist. Circumstances may alternatively require bridge=* on the road or ford=* on the crossing. None of these tags should be used unless the crossing type is known, despite warnings from some editing software that highway crossing waterway could be an error.
Bdentremont (talk) 03:14, 24 January 2024 (UTC)
- +1. Maybe add something like "Adding blindly guessed data to silence automatically detected errors is completely unhelpful." Mateusz Konieczny (talk) 16:34, 24 January 2024 (UTC)
Rendering support
Neither Potlatch2, nor Mapnik render tunnel=culvert as expected. What's the point? -- 17:31, May 10, 2011 User:Ponzu
- It's more descriptive than tunnel=yes, and captures a distinction people make in real life. The point of tagging is to capture those distinctions. I'm not particularly bothered about how it renders, but since you have expectations about that please consider raising a pair of tickets at https://trac.openstreetmap.org/newticket explaining what your expectations for rendering or editing are. Thanks! --achadwick 13:59, 21 January 2013 (UTC)
- (1) Tagging for the renderer, (2) it is no longer applicable (3) if you use some tagging getting support is more likely Mateusz Konieczny (talk) 10:39, 23 January 2024 (UTC)
Resolved: Mateusz Konieczny (talk) 10:39, 23 January 2024 (UTC)Culvert vs. tunnel=yes
What is the advantage of tagging tunnel=culvert over tunnel=yes? If a stream goes into a pipe for a while, what else is going to be if not a culvert of sorts? -- 17:31, May 10, 2011 User:Ponzu
- Greater descriptive power, and data consumers not having to draw the inference you mention. We don't just use road=yes or building=yes for example. I'd argue there are useful distinctions to be drawn too between "tunnels" and "culverts": assumed bore diameter, methods of construction, visibility in the landscape, relative position to transport infrastructure and relation to it, and rather importantly for human users of the map, what people call it and what it might look like on the ground. --achadwick 14:21, 21 January 2013 (UTC)
Removed advice to tag foot=yes if high enough for walkers
Removed the sentence
If the culvert is tall enough for pedestrians, add foot=yes.
because the advice is simply incorrect. foot=yes is a direct statement that there is an official, legally-enshrined (and verifiable) right of way; see Key:access. Do not use foot=yes to indicate that a culvert is high enough to allow pedestrians to walk through it.
We ought to have a way of encoding the height of a tunnel, but this is not the right way to do it. It should probably be covered more generally in Key:tunnel.
--achadwick 13:47, 21 January 2013 (UTC)
Resolved: Mateusz Konieczny (talk) 10:39, 23 January 2024 (UTC)"Tagging controversy" section
This section seems wholly uncontroversial now, with 43800 ways in the database comprising very nearly 10% of all ways tagged tunnel=*. I've moved it here since it's really historical discussion and the mailing list links and usage counts are worth keeping.
Tagging controversy
tunnel=culvert is how some people want you to tag culverts. Before this wiki page was created the tagging of culverts was converging towards culvert=yes. Which was documented a few days before the creation of this page because that tag was actually in use.
Tagging statistics of the 19th of jan 2012:
- man_made=culvert: 1 uses (
at least one).- waterway=culvert: 107 uses.
- tunnel=culvert: 18611 uses.
- culvert=yes: 2239 uses.
- and many tunnel=yes for culverts before this proposal
For the whole discussion: [1]
As of the 1st of February 2011, tunnel=culvert has become the most frequently used of these tags with 3656 uses. Current statistics are available on taginfo ([2], [3]).
Hope that isn't treading on anyone's toes; just cleaning up for the sake of presenting a concise, descriptive article for the tag. --achadwick 14:49, 21 January 2013 (UTC)
- No problem. At least we didn't get a tag with colon in it for this. --Cartinus (talk) 22:10, 1 February 2013 (UTC)
Cleanup remaining culvert=yes occurences ?
Recently stumbled upon some older waterways that were tagged "culvert=yes,layer=-1". Wondering if it would be a good idea to have them converted automatically to tunnel=culvert now? RicoZ (talk) 19:47, 16 November 2013 (UTC)
Get rid of "bridge=culvert" advice?
The page contains the sentence "In some countries, the culvert structure is directly reused as a bridge in which case prefer the tag bridge=culvert". However that was never approved afaics and appears rather confusing. Where should be bridge=culvert applied to - the upper, the lower or both ways? Description suggests both ways should be tagged as bridge=culvert and because they are a single structure they should also share a node or ways.
Should that be removed? RicoZ (talk) 14:05, 14 July 2015 (UTC)
- I came to the conclusion that this is more like a bridge than a culvert. That do you think? Description should be clarified, of course.--Jojo4u (talk) 14:16, 14 July 2015 (UTC)
- Hoping that someone who added or endorses this bit of description will voice his/her opinion and clarify, otherwise I would plainly remove it. There is also bridge=simple_brunnel which is in an early draft stage and might serve a similar purpose (as far as I can guess the original purpose of bridge=culvert).
DONE : the wiki pages have been changed quite some time ago and it was most likely a mapping accident.
If you encounter "fresh" occurrences of bridge=culvert try to clarify and correct the issue with the respective mapper. Editors/QA tools should generate warnings if someone tries to create them accidentally. RicoZ (talk) 08:49, 26 August 2018 (UTC)
Resolved: sounds resolved Mateusz Konieczny (talk) 10:48, 23 January 2024 (UTC)Culvert mapping use cases
What is the recommended approach for modeling culverts? The OSM editor is now displaying a warning when highways crossing waterways features are not connected. The editor also recommends "Highway crossing waterways should use bridges, tunnels or fords".
I have seen multiple cases:
1) The culvert is just a small pipe, trench channel, or very small concrete structure that does not extend or barely extends beyond the width of the highway. For example, there may be a small road, fire road, or even single-track trail that crosses a small creek. It seems overkill to map the culvert as a "ways" when the culvert is just a small pipe that helps to drain the water under a small road. Sometimes, the creek is so small that the path for the water is a U-shape trench channel installed across the trail/road. For these cases, what is the rationale for not recommending to use culverts on node elements? If we really have to model the culvert as a way, it seems it would have to be the smallest possible segment, but then why not model it as a node? I see a lot of OSM contributors seem to have avoided this problem by not connecting the "highway" and "waterway" features, but now the OSM editor is warning the user when these features are not connected. Bridges and tunnels are not supposed to be used on node elements. "Ford", on the other hand, is allowed to use a node element. Why not culvert?
2) The culvert is a narrow structure (e.g a pipe or concrete structure) that extends significantly beyond the width of the highway. In that case, presumably the culvert should be mapped as a segment ("way") along the waterway, not the highway. This would be unlike bridges, which are mapped as a segment along the highway. "narrow" is a bit subjective so let me give some examples: I have seen culverts that are made of a pipe whose diameter is less than 1 meter, and the engineers have decided that the pipe should extend beyond the sides of the highway by several meters. In other cases, the culvert is a narrow concrete structure whose opening is small (one or two meters), and it extends significantly beyond the sides of the highway; visually the culvert looks like a tunnel rather than a bridge.
3) The culvert is a large structure that does not (or barely) extends the width of the road. It has been built to let a large flow of water go through under the highway. For example, I have seen concrete structures that have multiple openings for water. In that case, should we map the culvert as a segment along the highway, similar to what is done for a bridge? This would make sense because from the perspective of the highway, there is a significant (more than ~10 meters) segment of the road that goes over the culvert.
4) What if the culvert is both large and extends significantly beyond the width of the highway?
https://wiki.openstreetmap.org/wiki/Key:tunnel https://wiki.openstreetmap.org/wiki/Key:bridge https://wiki.openstreetmap.org/wiki/Key:ford
- Regarding (1) there is Proposed features/Simple one node bridge which comes pretty close to your idea. Some years ago the idea was fiercely opposed because it was claimed that a road and waterway should not share nodes but I believe in the meantime many mappers gained the insight that a simple culvert is in many cases an integral part of the road construction and as such *should* share one or more nodes with the road. So the idea would seem acceptable by now - it just needs a new tag name because nobody can make any sense out of "simple brunnel" :) RicoZ (talk) 19:37, 5 May 2019 (UTC)
- RicoZ, I would not necessarily assume that the OSM community would be more receptive to your proposal today. For me, the main issue is semantics: A node between intersecting ways is used to expressed an at-grade crossing (that is, same elevation of the two intersecting features). This is why a ford can be mapped mapped as a shared node, whereas bridges and tunnels shouldn't.
- The meaning of a shared node is more varied than grade crossing. It has a special meaning when mapping "covered", lifts, waterway=dam and probably many other features. The grade crossing semantics applies to some special though widely used features like intersections of roads or confluence of creeks/rivers. RicoZ (talk) 20:34, 9 May 2019 (UTC)
- Srosset, the currently recommended approach doesn't really change depending on the size of the culvert: Split the waterway at the beginning and end of the culvert, and the relevant tags to that short new way. Whether the road or the waterway should be split depends on whether the construction is more like a "bridge" or a "tunnel". This can be a fuzzy distinction, the introduction has of Key:bridge has a few words acknowledging this and describing the usual rules of thumb. --Tordanik 20:00, 9 May 2019 (UTC)
Ditch still visible
A ditch passes under a road. The ditch is visible its entire length there under the road, via w:grating. How to tag? See also Talk:Key:covered#Grating. Jidanni (talk) 17:11, 21 August 2023 (UTC)
Resolved: duplicate of https://wiki.openstreetmap.org/wiki/Talk:Key:covered#Grating Mateusz Konieczny (talk) 10:49, 23 January 2024 (UTC)