User:Alfons234/Municipality
My view on municipality
Municipality as a concept
Following Jo Cassel's friendly advice, I try to outline my present view of the tag place=municipality here as coherently as possible.
This refers presently only to the situation here in Germany - more research and details must and will follow - but people have been (understandably!) unhappy with the fact that placenames of certain administrative entities do not appear on the map for going back to at least 2015. Related discussion threads on the forum: here(2022), here(2020), here(2018).
place=municipality first appeared on the map and in the wiki in 2009, followed by a steady (until 2020) and then sharp increase in uses. It was defined as "a single urban administrative division having corporate status" - almost the exact description of the places whose names are either missing on the map (Carto) or that have been forced onto it by misusing place=town. This misuse is believed to be a serious problem because it says "there is a town centre here" where there is no town centre. This corrupts the information contained in the database.
Collection of examples:
It is true that this tag logically duplicates boundary relations. However, not only can OSM's data model put up with redundancy but more importantly, it helps reflect and account for the fact that boundaries (not physical ones now, but those between concepts) can be blurry. Place nodes for cities, towns etc. are universally acknowledged (with some people preferring area over node), but our administrative entities may be viewed by the public as proper "places" to varying degrees, differing from place to place or in the course of time. Map_HeRo mentioned Wuppertal in this regard. Wuppertal, originally an artificial construct made up from the towns of Elberfeld and Barmen, is now a placename used by everybody. In contrast, people won't say they travel to "Hohe Börde" oder "Oberharz am Brocken" although they already use these names in their postal addresses, but this may change over time. So my view is that municipality is here to stay, with a perspective of its being used as a kind of "generic" placename tag, a bit like highway=path.
If this all is true, then the crux then lies in how the data is being used. Two prominent cases are (a) its rendering in Carto and (b) possible implications for routing engines.
The recent Github issue 4004 addressed the rendering in Carto in connection with boundaries. My instinct is to consider simply rendering place=municipality like place=town, independent from boundaries. But before doing so, it would be neceassary to look more closely at how place=municipality is used in comparison with place=town or place=village in other countries.
When it comes to routing engines, stopping the misuse of place=town can stop misdirections to useless geographical centres of administrative entities, but can replacing town with municipality also provide useful directions? Probably not, unless a rule for the position of municipality nodes were agreed on. Such an agreement (and implementing it anytime soon) seems out of reach. So this is a shortcoming.
Much more detailed thought on the wider framework and its implications has been given in several places by many people. Still, it is also legitimate to take on the single aspect of place=town misuses separately from the rest.
Implications of rendering municipality in Carto
Getting an overview of the tag in Germany
I assumed this to be a piece of cake after overpass only found 8 uses in Lower Saxony, but this low number seems to be the exception rather than the rule: more than 8000 such nodes in Germany!
The majority of them are what I expected them to be: new, often fancy names given to a group of separate villages. For these, the simple and perfect solution is, as far as I can see, using exactly the same rendering style as for place=town.
However, the situation can also be different:
- The area of the municipality may be really huge, such as Nordwestuckermark. This is not such a big problem because along with huge areas there will be a low density of features, so you can still find the label quickly.
- The newly created municipality may not have been given a new name, but the name of the largest town within the municipality is now used for the whole of the municipality as well as for its original namesake. Example: Körle. In these cases, using the same rendering style as for town is problematic, because the name will appear either twice or not at all, depending on zoom level. Deleting the town node is obviously not acceptable. So the municipality label either has to look different from the town label, or tagging guidelines (say, "Don't create a municipality label if there is a town label of the same name nearby.") will be influenced by the rendering outcome, which turns the idea of tagging and rendering upside down to some extent.
- The towns may really have grown together, so the original towns come to "feel" like suburbs within the new municipality, as, possibly in Ehrenkirchen. This is less problematic again, because in this case it will be perfectly legitimate to rely on the town node. The municipality label is abundant, because it duplicates the admin_level=8 boundary relation.
So at present, having looked at Germany only, the issue boils down to three alternatives to address the challenge that places like Körle present:
- Try to convince the community of problematic and most likely controversial tagging advice (proposal) in connection with the attempt to have municipality rendered?
- Accept that sometimes the map will show the same name twice, in exactly the same typeface?
- Devise a new rendering style for municipality, i.e. think about zoom levels (earlier than town?), maybe grey, in italics, a little larger than town...?
Does that cover all the possible implications? For that we need to look at other countries.