This is a single building that offers lodging, restaurant(s?), a brewery, a bar, and possibly more, all seemingly under the same name. Normally I would be comfortable tagging multiple amenities as separate nodes with the same name if it were just two, but this seems excessive and possibly a burden to data consumers. I’m also not exactly sure where inside the building each amenity is, so it’s a little more complicated for me than just placing each one where it belongs. As currently mapped, the building has a slightly incorrect name which I could easily fix, but it’s also just tagged as “motel“, which is misleading. One specific question: if I were to split this out into multiple nodes, would the building carry a name and a tag set or would I leave the building alone in that case rather than deciding which set of tags the building gets? For the rest: I was hoping more experienced members could point me towards relevant mapping guidelines and maybe even propose a few examples of how I would tag this based on those.
I’m not an expert, but I’d suggest tagging the building itself with the name, website, address, etc. (common attributes), and then adding individual points for the individual aspects of the business (accommodation, brewery, event venue, etc.).
This wiki section seems to suggest it’s OK: One feature, one OSM element - OpenStreetMap Wiki
Either “multiple points for separate amenities” or “one point for the main amenity with qualifying properties” might be OK.
To give an example of the second, consider an amenity=pub with accommodation=yes. Here’s an example, and here is a map (one of mine) that shows accommodation as a blue roof over the underlying pub symbol. Similarly microbrewery can show like this, and events_venue could also be shown differently if desired (although I don’t have an example of that).
The alternative of multiple POIs, like here or here, also works.
It depends whether you think they are “all one thing” or not, I guess.
When you say individual points, do you mean just tagging the building with the name and common stuff and then putting unnamed points inside with the individual features? I can see both sides: if only the main building is named, there’s only a single search result and it’s still obvious that a restaurant point inside it belongs to it. But anyone trying to find restaurants near them wouldn’t see it (or would see it without a name) if the main building was say, a hotel (or just a building with no amenities).
The individual points could be named, such as “Woodstock Inn Wedding Venue”.
As @SomeoneElse said, you can tag certain things with multiple attribues. So if the main building were tagged as a hotel, you could add accommodation=yes to cover that aspect. Then the individual points would just be for the really different parts (like the brewery and events venue).
From looking at their website, I would draw the area of the whole site (including buildings, driveways, etc.) as amenity=events_venue (name, address, etc.), and map inside the brewery and the accomodation, pub etc. You can use “level” if you want to indicate a specific building floor for a feature.
You can use areas for the features, or nodes. Areas convey additional information and allow for intersection with other areas and nodes (e.g. what is inside what, size), while nodes are easier and faster, and sometimes you don’t know the shape and extent of the area or you don’t care.
Features do not have to be tied to buildings, rather buildings can be seen as features on their own, so ideally they are separated. You can overcome this problem also with properties like accommodation=yes but in this case it is not common and can easily be overlooked: accommodation | Keys | OpenStreetMap Taginfo
Still if the accomodation was seen as a service of something else, this would probably be the events_venue and not the building. Or the brewery spotting an events_venue and accommodation. As @SomeoneElse said, it all depends how you see the “structure”.
That sounds like a decent solution. One last pedantic point that’s stopped me from doing that so far: best practice is typically naming things as they appear on the ground, and if there’s no signs on the ground naming these like this, that would technically violate that right? I’m probably taking things too literally/seriously, but I’m just not sure where the line is between descriptive and technically inaccurate.
On the ground can also mean that people on the ground call something with a certain name. It doesn’t have to be a physical sign.
no, physical signs are not needed
maybe we should update wiki to clarify that…