Proposal:Rivers Classification
| Rivers Classification | |
|---|---|
| Proposal status: | Rejected (inactive) |
| Proposed by: | Zverik |
| Tagging: | river=* |
| Applies to: | 
|
| Definition: | Subjective classification of rivers |
| Statistics: |
|
| Draft started: | 2017-08-29 |
| RFC start: | 2017-08-29 |
| Vote start: | 2017-10-16 |
| Vote end: | 2017-10-24 |
Proposal
Let's have a subjective river classification, so we could tell a big river from a small one.
It could be based on measurable (google-able) parameters, like length or width. The attribute would be useful for making low-zoom maps: for simplification, for rendering and for querying. It should be understandable to casual mappers, and should not affect any other tags.
Tagging
The following tags are proposed, along with the way to separate these:
- river=major — a major river, more than 1500 km in length, or with basin of more than 300 000 km², or that contains any major rivers in its drainage basin. Basically, the first 91 rivers from wikipedia:List of rivers by length should be tagged with this.
- river=big — a big river, more than 300 km in length, or with basin of more than 15000 km², or that contains any big rivers in its drainage basin.
- river=small — all other rivers. This is considered a default value: when there is no river=* tag, a river is considered to be small.
These should be used together with waterway=river. Smaller waterways should still be tagged with waterway=stream, thus making it a fourth category of waterways.
Note that rivers form a network, and the size must only increase: streams flow into small rivers that flow into big rivers that flow into major rivers that flow into big lakes and oceans. Some of these steps may be omitted, but never a river flows into a river with lesser size. Lakes inbetween do not matter.
The threshold for big rivers could differ by continent or by country: for example, in UK it could be lowered to 200 km, and in US — increased to 500 km, to omit many smaller rivers. This proposal introduces the fixed value to serve as a rule-of-thumb.
Rationale
Copying from the alternative proposal:
- OSM has only one way of general categorization waterways - streams and rivers (other than canals, drains and other well defined entities). It's very handy at the small scale, but at the large scale all the rivers are basically the same in our database. It should be possible to have general categories like we do with roads network, so we could for example render world map with only the most important rivers.
The obvious thought is to calculate the size of a river automatically or use a waterway relation. But the first way is complex and would benefit only a few mapping styles that are maintained by skilled developers. The second way is better, and you could use a distance=* tag from a relation. But these relations are mapped inconsistently, and even if you could maintain a hundred relations for major rivers, I doubt all the big rivers would have their relations mapped in at least a decade, without breaking any of these.
Thus I think that tagging river ways is the better option. But it could involve mass-tagging after the proposal is accepted.
Examples
- Ways in River Volga (3645 km) and Rio Japurá (2615 km) should be tagged with river=major.
- Ways in River Wisła (1047 km) and River Le Tarn (381 km) should be tagged with river=big.
- Ways in Neva River should be tagged with river=big, since despite the small length (only 74 km), this river has an enourmous drainage basin, 281 thousand km². It may even be considered for a major class, to be shown on all maps that have Lake Ladoga.
- Ways in River Yauza (48 km) and River Nene (161 km) should be tagged with river=small or not tagged with this key.
Applies to
The attribute should be used only on ways tagged waterway=river.
If a river has a waterway relation, do tag it the same. You still need to tag the main stream ways.
Rendering
The value of river=* would basically tell on which zoom level the river should be rendered.
Existing tag values
Taginfo shows around 150 uses of the river tag on ways. Most common values are river=yes, river=pond and river=riverbed. I assume we can safely disregard these.
Comments
Please comment on the discussion page.
Voting
I oppose this proposal. The descriptions still seem too arbitrary, and a mix of importance and size. Seasonal rivers could be major at some times of the year, and dry at other times of the year. The same is true for streams, in one location local people would call it a river, and in another location it would be called a stream. --Palolo (talk) 17:00, 16 October 2017 (UTC)
I approve this proposal. Still looks sane. --Zverik (talk) 10:17, 16 October 2017 (UTC)- I approve this proposal. Wondering about what falls in between small and big, knowing that that is not easy to say. Pander (talk) 10:28, 16 October 2017 (UTC)
I have comments but abstain from voting on this proposal. Conflicting definition of =major "Basically, the first 91 rivers from wikipedia:List of rivers by length should be tagged with this." vs "It may even be considered for a major class, to be shown on all maps that have Lake Ladoga.". Also, see Talk:Proposed_features/Rivers_Classification#An_overall_review. --Mateusz Konieczny (talk) 11:11, 16 October 2017 (UTC)- I oppose this proposal. I recommended to remove the opening statement about this being a "subjective" classification as it will send the wrong signal. This proposal is not more subjective than, for example, highway tagging, and it shouldn't claim to be subjective. My recommendation has been ignored, therefore I sulkingly vote no. --Woodpeck (talk) 13:00, 16 October 2017 (UTC)
- I have comments but abstain from voting on this proposal. I'm participating in the voting because I think the idea has some merit. However, I'm not sure it's practical. Classifying rivers by major, minor, etc. seems somewhat arbitrary, plus it seems hard to verify. I know for one that I have neither the inclination nor resources to determine the necessary attributes in order to correctly tag a river, according to the proposed scheme. If someone else wants to do it, I will probably accept whatever they decide. I personally don't have much interest in classifying rivers, because I only am concerned with the part of OSM I personally visit and make use of, but I can understand how it would be important to some folks. --Dr Centerline (talk) 14:40, 16 October 2017 (UTC)
- I oppose this proposal. I agree that river data by OSM is not useable for lower zoom levels. But this proposal does not help. Every data consumer has a different opinion which rivers are important. Your scale is too rough-grained to be useful (see the useless distinction between usage=main and usage=branch for railways. If you want to render a map of western Europe (longitude -10 to +15° from Greenwich), you don't care how long River Wolga is. Even the length of River Danube does not matter although it is on the map because its section above Passau is "small". Better map different features like use by large ships (observerable, size is often written on the ship) and of course the width. --Nakaner (talk) 20:26, 16 October 2017 (UTC)
- I oppose this proposal. unclear definition --Waldhans (talk) 20:45, 16 October 2017 (UTC)
- I oppose this proposal. I like the proposal, but I don't think its complete yet. Because it's subjective, it's likely to cause edit wars. Nominatim has an interesting approach. They look to see if wikipedia has an entry, if it does, it gets bumped up. The proposal is also close to adding in a "notability" requirement, much like wikipedia. I don't think that belongs in OSM. Glassman (talk) 13:11, 17 October 2017 (UTC)
- I oppose this proposal. I think the only real goal of the proposal is to make Neva River "bigger" on some map Smollett (talk) 14:07, 17 October 2017 (UTC)
- I oppose this proposal. Please read the Verifiability article. --Ignaciolep (talk) 18:10, 17 October 2017 (UTC)
- I oppose this proposal. I'm convinced by Imagico's arguments. Rather than a short-term subjective solution, what is needed is long term investment in tools to calculate characteristics of rivers from the data itself, giving the map designer the freedom/responsibility to determine importance --JoeG (talk) 05:29, 18 October 2017 (UTC)
- I oppose this proposal. Adavidson (talk) 03:12, 19 October 2017 (UTC)
- I oppose this proposal. The length might be prone to be incomplete already and for most of the watersheds in eastern asia we would have to look at external sources as especially medium and small sized rivers are incomplete and occasionally cross-connected by canals (mainland china). In my opinion boat=yes, waterway=riverbank and maybe width= are the better tools to determine importance and render priority. --Helmchen42 (talk) 10:30, 22 October 2017 (UTC)
- I oppose this proposal. see remarks above --Hfst (talk) 13:43, 22 October 2017 (UTC)
- I oppose this proposal. --Joto (talk) 08:15, 23 October 2017 (UTC)
- I oppose this proposal. --Soldier Boy (talk) 18:57, 23 October 2017 (UTC)
- I oppose this proposal. It seems to be a subjective matter. The expected information can be achieved by the existing tagging scheme. --BastiNB (talk) 21:25, 22 October 2017 (UTC)
The author ended the voting prematurely, since it's clear it won't get accepted. Sorry folks.
See Also
- waterway=river
- Relation:waterway
- Proposed features/Waterways classification (and this section in the discussion)