Talk:Proposed features/Inlets proposal

From OpenStreetMap Wiki
Jump to navigation Jump to search

Inlet=Screen Question

Resolved: Values has been adjusted and classification document link has been added to proposal Fanfouer (talk) 22:26, 5 September 2021 (UTC)

The picture shown for the inlet=screen on the page here, to me, doesn't really represent what a stormwater screen usually means. What's shown is typically called a stormwater drain cover, or grate. Clarification and additional notation on stormwater inlets. See page 15 (on viewer) or 7-13 (page-number notation). --IanVG (talk) 16:45, 21 December 2020 (UTC)

Hi @IanVG:. Thank you for this really useful documentation I now lined in the see also section at the bottom of the proposal. That's right, my screen value was a bit too wide and some additional values have been added in the table. Feel free to add examples if you have some, I'll look in my neighbourhood as well. Fanfouer (talk) 22:14, 25 February 2021 (UTC)
Great @Fanfouer:, good contributions! Coincidentally, my father has taken pictures of storm grates from all over the world, as he wants to write a book on them (go figure), so hopefully soon I can upload some more examples of stormwater drains onto this proposal. I also wanted to ask if you think it's a good idea to add the tags inlet=combination and inlet=slot (the latter potentially as a way too). They are referred to and described on page 7.27. Let me know what you think. I wanted to make sure that perhaps one of the existing tag-values doesn't already describe these features. --IanVG (talk) 22:55, 6 March 2021 (UTC)

Single Tag Value for Way

@Fanfouer: I think there is one exception that could merit using this tag for ways (and not just nodes). A slotted drain inlet (I don't have a picture, but it is on page 7-27 of the Urban Drainage and Urban Flood Control Report). It seems that a tag designated for either nodes or ways would allow this kind of feature to be mapped.

Thoughts? --IanVG (talk) 01:30, 26 July 2021 (UTC)

Given problem is that it would encourage to use it on all pipelines and tunnels ways as well and prevent easy quality control by checking the geometry. I'd like to define inlets/outlets as opened (punctual) extremities on closed ducts like tunnels or pipelines.
Furthermore, I think slot drains aren't consistent with pipe definition since they have an open side in the length direction (while a pipe is only open at its extremities, aren't you?). Here we should consider a completely different feature, unless gutters or any canal will also be defined as a pipe with extra exaggeration.
However, this is a good object to document in OSM. We'll find a way Fanfouer (talk) 21:05, 27 July 2021 (UTC)

Kerb not curb

Resolved: Kerb is now used

Please use British English in tags, particularly for consistency as barrier=kerb and kerb=* are already in widespread use. SK53 (talk) 09:42, 25 September 2021 (UTC)

Done. I wasn't aware of this difference, thank you Fanfouer (talk) 11:53, 25 September 2021 (UTC)

inlet=tulip Is not really a thing

Resolved: Tulip was moved to bell_mouth

A Google search for "tulip spillway" gives me 39 results, "bell mouth spillway" 4,600, and "morning glory spillway" 15,000 results. I'd suggest using one of the two later options. Adavidson (talk) 07:40, 17 October 2021 (UTC)

Is "tulip" a literal translation from French? Adavidson (talk) 08:52, 17 October 2021 (UTC)

Indeed that was a literal translation from French. I've changed it for bell_mouth with a link to wikipedia page. The same for outlets proposal. Thank you Fanfouer (talk) 13:04, 17 October 2021 (UTC)

@Stefanct: during first vote: The bell_mouth value is technically OK but from an internationalization standpoint I think something like funnel would be way easier to understand for non-native speakers and would be more generically applicable (which might not be desirable though). The most important property beside its physical appearance is missing: when is it active - always, by design in overflow conditions, after mechanical intervention (manually or by some sensor/actuator control)? Stefanct (talk) 11:14, 18 December 2021 (UTC)

Depending on the vote's issue, I may consider funnel as a possible refinement.
The point of the proposal is to classify inlets according to their shape/design. It is expected to have further and several proposals to give details about each one. It wasn't the point to describe what the precise operational status could be for now. It doesn't mean it's not important, it only means it wasn't the right time to discuss it. Fanfouer (talk) 20:46, 19 December 2021 (UTC)


Resolved: grate can be used on any location, not only pavements

"usually located on pavements"? In my experience it is rather located on road which is lower that pavements. Otherwise road would become flooded Mateusz Konieczny (talk) 08:05, 17 October 2021 (UTC)

Also, why it is listed in "Culverts or tunnels" section? We are not mapping stormwater collection tunnels, right? Mateusz Konieczny (talk) 08:06, 17 October 2021 (UTC)
Indeed it wasn't relevant to mention pavements only, I've changed the description for simpler streets. Fanfouer (talk) 13:09, 17 October 2021 (UTC)


Resolved: inlet=yes has been added to possible values

If I dont see the inlet directly to determine the type, can I map inlet=yes? Nielkrokodil (talk) 13:51, 24 October 2021 (UTC)

You'll always be able to do so, as per Any_tags_you_like, just like any undocumented situation. It's not necessary to add it in this particular proposal. Do you already have found such inlets you weren't able to qualify? Fanfouer (talk) 14:35, 24 October 2021 (UTC)
I know "any tags you like" but not everybody does. I would prefer having this option mentioned in the wiki for mappers who maybe wouldn't map it if they are unsure of the type.
No, I haven't found an inlet of other type. I wanted to map one some years ago but didn't because I found no fitting tag. I need to revisit the site to check the type, because I don't remember it any more. Until then I will set inlet=yes. Nielkrokodil (talk) 15:02, 24 October 2021 (UTC)
Do you think the same would be useful for Proposed_features/Outlets_proposal? Fanfouer (talk) 18:25, 24 October 2021 (UTC)
Hm, I don't know. I cant think of such a case right now.
I now recall the situation better where I wanted the inlet=yes
I mapped a stream which was led underground once it reached a residential area. I did not know where it continued underground, so I just "stopped" the stream there.
And then I got a JOSM warning of an unconnected stream which I wanted to resolve. I believe the inlet tag will do that.
(This particular inlet is well hidden behind bushes, so it is not easy to see which kind of inlet it is)
This is a bit similar to the noexit=yes tag on highways to say "this really is the end here, it is no mapping mistake" Nielkrokodil (talk) 20:19, 24 October 2021 (UTC)
Ok that's a good point regarding validation, I've added inlet=yes to the proposal Fanfouer (talk) 11:01, 26 October 2021 (UTC)

Thank you. Do we have to fear someone armchair-mapping every stream-to-culvert node as inlet=yes? Am I right that such a node, without any tag itself, implies inlet=yes and therefore does not need it?

Nielkrokodil (talk) 18:43, 28 October 2021 (UTC)

What about kerb grate without opening in kerb?

I have seen ones similar to - without opening at all in kerb itself Mateusz Konieczny (talk) 12:26, 15 November 2021 (UTC)

That would be simpler inlet=grate if no kerb opening is visible. Fanfouer (talk) 13:29, 17 November 2021 (UTC)
It occurs to me separating the two will be clearer,
---- Kovposch (talk) 13:08, 14 December 2021 (UTC)
I think you're right and thank you for this suggestion. I'm not used to tag barrier=kerb on nodes, isn't it relevant on ways only?
Depending on the result of the vote, kerb:inlet=* is likely to be integrated or will definitely find its way in a further proposal Fanfouer (talk) 15:08, 14 December 2021 (UTC)

How it would be tagged?

What would be full tagging for objects like ?

Are you proposing to map individual raiwater collection points on roads like this one? It seems a bit too much for me Mateusz Konieczny (talk) 12:28, 15 November 2021 (UTC)

Yes, as many nodes as actual rain water inlets. Each of them can have different capabilities and properties.
This is a tagging model, no one is forced to map every node. Defining inlets properties on the road's way sounds not valuable as well as we'll not be able to know where each inlet actually is. Fanfouer (talk) 13:29, 17 November 2021 (UTC)
Would be more useful and less demanding if it can show a road section or kerb line has a certain inlet design, similar to passing_places=* and lit=* vs highway=passing_place and highway=street_lamp. I forgot why man_made=inlet is not used, otherwise this can be added here. ---- Kovposch (talk) 13:25, 14 December 2021 (UTC)
I'm not sure about usefulness if we come to document every inlet with individual properties (flow rate, dimensions, material, usage). This proposal is set up from a water management perspective and always using node is done on purpose. We had the same discussion regarding road with utility poles beside, a road attribute had been asked to tell if the road was along poles. Despite it's more demanding, it's more usable to use man_made=utility_pole on individual nodes Fanfouer (talk) 14:40, 14 December 2021 (UTC)

Proposal purpose

In its current state I don't see the use cases this tagging is meant to serve. It appears to simply allow mappers to add various small features that are more or less related to drain water management to the database. That's fine in the sense that it channels (sic! ;) mappers into using consistent tags but from a data user point of view I am a bit puzzled how this helps anybody. The values don't make a lot of sense to me as they are overly specific without a good reason. That's especially true for a proposal that advertises its extensibility towards other uses than (drain) water. Stefanct (talk) 11:14, 18 December 2021 (UTC)

The proposal is intended to increase ability to describe places where fluids (not only water in any mean) are collected and directed into a closed duct (tunnel, pipeline, culvert, whatever). Not more.
They are usually found in hydropower facilities, industrial plants or even in the streets with rain water.
I don't get what means overly specific according to you, what classification would you make regarding those features? Fanfouer (talk) 20:46, 19 December 2021 (UTC)
Example: you won't find a kerb_grate anywhere but on streets for wastewater. Similar for other values. While you write (here, in the proposal and further down as well) that the proposal covers other applications too, that's a claim I don't buy. Of course there are inlets in other applications but that by itself does not make this proposal universal by any means and the very specific values assert that it is not. Your first sentence here btw agrees with my assessment: the purpose of the proposal is to enable mappers to add stuff but has not considered the data users end. --Stefanct (talk) 21:30, 19 December 2021 (UTC)
Asserting a given key (this proposal is about a key) covers several kinds of fluids isn't incompatible with values that cover one kind of those fluids. highway=* doesn't cover car paths only even if some of its values are specific to cars, right?
Regarding data users, do you have some existing practices in mind I could integrate? To me, we start here out of nothing in inlets mapping, just like outlet=*s. Usages will rise when data will be available and then refined is necessary, just like many fields on OSM Fanfouer (talk) 21:42, 19 December 2021 (UTC)


Apparently strainers are mapped according to their orientation and shape: horizontal=grates, vertical-ish=screen. The kerb-specific values are unnecessary because they should be attached to a barrier=kerb and could use the direct or strainer-specific value. Stefanct (talk) 11:14, 18 December 2021 (UTC)

I agree this should be stated more precisely in definitions: screens catch debris and grates prevent humans, animals, vehicles to fall in holes they may find on their path. A grate supports what is supposed to pass over it. A screen only retains something in a fluid flow. That's a big difference, you can't swap them as they are designed for different usage
It's a valid point regarding barrier=kerb, as mentioned by Kovposch upside. A better integration depends on the vote result. Another proposal may be necessary to go further on kerb inlets classification. Fanfouer (talk) 20:46, 19 December 2021 (UTC)

Overlap with other features

The proposal also focus only on closed water infrastructure. Stefanct (talk) 11:14, 18 December 2021 (UTC)

No, the proposal is directed toward inlets admitting any kind of fluid, according to proposal and rationale chapters.
It's true there are only examples showing water, we miss other fluid situations and it could be completed in the future, according to the proposed definition. Fanfouer (talk) 20:46, 19 December 2021 (UTC)

I know many cases where strainers are installed in open canals for agricultural use or in alpine streams (for protection known as torrent control or in German: Wildbachverbauung, e.g., - as of now I would interpret the proposal text to not apply to open streams. *If* they would apply then there is clearly an overlap with waterway=lock_gate and waterway=sluice_gate that should be addressed: is it a gate or an inlet=valve or both? :) Stefanct (talk) 11:14, 18 December 2021 (UTC)

The proposal only focuses on closed ducts/open air transitions. Despite it may be useful to define appropriate tagging for screens/strainers in every situation, inlet=* doesn't relate to open air features.
Note that waterway=sluice_gate is part of values we should avoid regarding waterways. A sluice gate isn't an actual water way, despite related to it. Fanfouer (talk) 20:46, 19 December 2021 (UTC)
Neither is a lock_gate an actual waterway, or milestones and traffic signs actual ways but still we use highway=milestone and highway=give_way. I don't know what you want to say. --Stefanct (talk) 21:36, 19 December 2021 (UTC)
A refinement of waterway=* is ongoing since 2018 to make it more focused on waterways. It took time to make it live, we go step by step. An existing practice isn't necessary the best one, it's a general point about whole OSM tagging. More precisely here, I would certainly not ever define waterway=inlet just because some inlets admit water sometimes Fanfouer (talk) 21:45, 19 December 2021 (UTC)

And the valve value has another problem/overlap that was not mentioned in the proposal: pipeline=valve. It might make more sense to split this proposal into values related/attached to pipeline=*, waterway=*, highway=* (or maybe barrier=*) keys IMHO. Stefanct (talk) 11:14, 18 December 2021 (UTC)

It should lead to an equivalent classification regarding inlets. Inlets occur on many different features: pipelines, tunnels. They could contain many fluids (waterway is only one possibility among other). How relevant would be several proposals defining what screens are regarding water, oil, sewage, whatever in different documents? Fanfouer (talk) 20:46, 19 December 2021 (UTC)