This is e.g. Abteiland or Falkensteiner Vorwald. But it’s many many more that are not mapped. But this question is also for Rheintal which is the valley of the river Rhein. Yes, I can search for the Rhein and know where the valley is, but shouldn’t the valley explicitely be mapped since it’s different from the river?
Sorry if tbis is in the wrong topic but I don’t know how to put it in the right topic
geographic regions typically aren’t well defined (in terms of cm-precision, which is the only possibility the current established OpenStreetMap data model offers). Unless they are defined by features like coastlines. As a result, these approximate regions, drawn as if they were representing precisely delimited polygons, are very confusing (and questionable) in high zoom levels. E.g. it is clear that a valley is around a waterway (or former or intermittent waterway, save few exceptions that may or not fall into the valley definition), but how high does it extend? Or regions named after geological features, how precise do you want to conduct the geological surveys?
So from the perspective of someone wanting them in a map, it is typically sufficient to have a rough idea (size, orientation/shape) so they can decide when and how to draw a label, but they won’t need or wouldn’t want a polygon taken literally at an 1:1 (or 1:5000) scale, especially for large features. In addition, these features if mapped with OpenStreetMap methods (often relations with pre-existent members that are very precisely drawn for different reasons) lead to very “heavy” and expensive geometries, for very little benefit.
My own conclusion was to refrain from mapping these things at all in OpenStreetMap, and starting a side project where we could collaboratively develop a parallel dataset that could be mixed with OpenStreetMap when producing maps.
license is cc-0 and pull requests (e.g. you have locally drawn a region or two that you know) are welcome: GitHub - dieterdreist/OpenGeographyRegions: A project to create a free, global, crowdsourced, multilingual dataset of geographic regions
I had a look into your repo, thx for sharing.
I fully agree with the precision problematic but that shouldn’t prevent osm from including information.
A river valley should be mapped according to wosm along the river or the center of it. This is sufficient information to map the valley properly. It doesn’t matter how the valley really extents, the center is enough.
To my understanding the same holds true for other areas, like regions, even though there is no exact boundary. A label would be sufficient if it has no exact boundary. If a boundary is available, then you can map it butit’snot a neccessity.
Your repo contains a good amount of regions that are currently not in osm like the Marlbourough Sounds but it also duplicates regions like Antacrtica. I am not sure I understand why Marlborough Sounds shouldn’t be mapped in osm like it is specified in your repo? If it’s the best approximation we have, why should we omit the information in the general dataset? And if it’s better to create a different dataset, shouldn’t the information be left out of osm?
I understand that deduplication can in theory become an issue, but I would not make the datasets interdependent (simple argument could be that it would be required to be ODbL if it is strongly tied to OpenStreetMap?).
So you are not asking for an (exact) “area” in osm but a label, hm?
OSM does not have the type “way” as labels, only “node” for a city etc.
There is a relation for the river Rhein with “name”=“Rhein”. In overview zoom levels, the name is not rendered.
You propose to add something like “region:name”=“Rheintal”. And on overview zoom levels, the label could be rendered, do you?
I think, for manual paper maps (not OSM) placing such region labels is difficult as they may hide other details. So they may be moved a bit, something an OSM renderer will not be able to do. City labels do have the same problem, were to set the label node.
Thanks for your reply @karlos
There are some areas mapped, but only very few. It’s not even that important that it is rendered on the map. It’s about someone telling you he’s going on holiday into the xy area. You don’t know where it is, so you can either type it into your preferred search engine, or go directly to the map.
A search engine will provide you the wikipedia result for the area. But it’s different for map applications. If they only show you results from osm (in this case), you’ll find nothing. That is not useful. In the best case, you’ll directly see where about it is, where the boundaries are (if known) and an excerpt of wikipedia with a link to the full article. Currently this information is missing almost completely.
The example of the Rhein is relevant because if you search for the Rheintal, you’ll have no result. You need to know if the valley is formed by a river or not and if so, you need to leave out the valley part in the name in order to find something. If the valley is not formed by a river you can’t leave the valley part out of the name.
I am not sure if I am proposing a new tag since there is already a tag place=region according to wosm. I am wondering why regions are not mapped and propose that we start mapping them. It’s missing information.
The tag place=region is not only existing in the wiki but already in use althoug not very common as it appears. If you search for well known regions like “Schwarzwald” or “Schwäbische Alb” you will well find those in OSM carto although the names are not rendered.
Other world wide known regions like the Provence in France apparently have not yet been mapped (when searching at least I can’t find an appropriate result rather than hundreds of villages, bakeries, houses, bus stops and the like containing “provence”).
So from my point of view everybody can feel free to go ahead with mapping such regions accepting the fact that the names will not be rendered in carto by reasons already explained in earlier comments.
oh yes, lots of Schwarzwald all over Germany and Austria 
there is the region after a lot of scrolling
? I don’t really understand what you mean. The Schwarzwald region ist mapped as a relation with roughly estimated border. The marker (after searching in carto) is placed well in the center of that region and it remains there regardless any scrolling, at least on my screen.
Schwarzwald is there, but not the sillicon valley, champagne or provence
i can’t find kings cross or just the cross in sydney either
A map rendering region names is opentopomap:
Will they get deleted if I add them?
It rather depends on what you add, and what you add them as. For example with Silicon Valley, that’s something of a historical umbrella term, and much of the activity it was previously renowned for has moved up the road or elsewhere. Before doing anything I’d definitely discuss it with the local (very active!) OSM community.
@SomeoneElse thanks! How do I get in contact with them? By looking who’s active over there and sending them a message?
The OSM US community is most active on OSM US’ Slack instance. The link for that is https://slack.openstreetmap.us/ , and in there there is a “#local-california” channel. There is also a talk-us mailing list.
Thanks! I’ll give it a shot
Thanks for reaching out to the U.S. community on Slack. For the benefit of those who aren’t there, Silicon Valley is already mapped as a place=region node following this talk-us-sfbay discussion. You can find it by searching Nominatim, but not by looking at most rendered maps, because place=region carries too little information about geographic scope to be labeled reliably.
It’s very difficult to decide where to place the Silicon Valley node, because it isn’t a strictly geographic term. Locals primarily use “Silicon Valley” to refer to the technology industry and only use it as a place name for marketing reasons, like how the City of San José bills itself as the “Capital of Silicon Valley” (admin_centre, anyone?). We use other names like “South Bay”, “San Francisco Peninsula”, and “Santa Clara Valley” to refer to a more specific geography. The place=region node ended up at a historical site, a garage where the West Coast electronics industry got its start, but most tech companies are further to the south these days, closer to San José.
The popular understanding of “Silicon Valley” differs outside of the San Francisco Bay Area. So our local POI import is named after Santa Clara County, but when I promote it to the broader OSM community as a case study in POI imports, I call it the “Silicon Valley POI import”, because people are more likely to know what that means, even though we’re importing far more than tech company headquarters.
Not every regional moniker is so fraught with ambiguity, but the problem of defining the region probably dissuades many people from mapping the regions they’re familiar with. Neighborhood names are similar, with some that are well-defined, others amorphous, and still others that are difficult to associate with any center coordinate.