Just wondering how the community feels on tagging historic ships (and submarines) as buildings. For clarity, these are ships that are now permanently moored and are often tourist attractions.
The historic=ship wiki documentation used to say not to tag historic ships as buildings. It now says the opposite. The wiki pages for historic=locomotive and historic=aircraft still explicitly say not to tag with building=*.
My concern is that this tagging is simply tagging for the renderer. The historic key is not rendered by Carto, whereas buildings obviously are.
It also seems a bit of a misnomer. A ship isn’t normally considered a building. Under the OSM definition a building is a “man-made structure with a roof, standing more or less permanently in one place”. Most dictionaries are stricter than this even, requiring walls as a well as a roof.
However, we do use building=houseboat and a permanently moored tourist attraction does kind of act as a building.
Thoughts? If we’re happy with using building=ship, should we also allow the same for historic, permanently located, trains and aircraft?
I agree. But vast majority of ships is not mappable in OSM as they keep moving.
And many permanently moored ones, mappable in OSM, are a permanent building-like structure. For closely related example lets take nonhistoric restaurant barge Way History: 702120324 | OpenStreetMap
It got moved only during very large flood and definitely does not qualify for historic=ship that is even less fitting than building=boat
The same goes for many museum ships.
Most are passing “man-made structure with a roof, standing more or less permanently in one place” and maybe even wall requirement, depending on how you define “wall”.
for me at least all of them are farther away from buildingness than ships/boats but not sure is it shared in general.
Or take Plik:ORP Fala.jpg – Wikipedia, wolna encyklopedia that used to be permanently moored in the same location for several years. Not really sure is building=ship really inaccurate, though I am holding strongly to that.
Is a decommissioned warship really a building-like structure? It seems like quite a stretch to me. Though I do agree your barge example is much closer to a building.
The same goes for permanently stationed: houseboats, camper vans, wagons used as buildings, containers used as buildings and so on.
Curiously, I would not apply it to such derelict objects or not used as buildings. Not sure is there any consistency here or is my classification not making any sense.
According to my knowledge constructions which can be moved by a build-in engine are not considered to be buildings but other structures (we would tag those usually as man_made).
A self propelled boat used for living purposes would be considered as a boat or ship and not a building therefore, whereas a swimming house without any self propelling device (but might be pulled away by a tug boat) represents a building according to this definition.
Presuming that would lead to the classification of land based structures like mobile homes, containers, caravans, railway carriages etc. as buildings if permanently located in one place and used for any kind of inhabitation as long as they are not self propelled and need towing vehicles to be moved to another place.
I do not know if this definition has some international backing but it makes sense to me as otherwise very difficult to say if a swimming structure could be a building or not.
floating buildings may not qualify as buildings in many jurisdictions, but I don’t see a big problem mapping them with building =* if they seem to be at the same place for longer periods. Definitely merits additional tagging though. I found a “building:foundation” tag, “floating” could be a first level value? https://taginfo.openstreetmap.org/keys/building%3Afoundation
The Delta King in Sacramento, California is another interesting example. It’s currently a restaurant/hotal and is being tagged as a building, without indication that it’s also a ship. I’m not really sure if that’s “correct” or not, but it would be good if there was some indication in the tagging that it’s a steamboat. I don’t think a ship stops being a ship just because it’s permanently moored. It looks like there’s 563 uses of building=ship. So maybe that’s the answer. I don’t know though
(As a side to that building=train seems fine, but building=aircraft is a bit of a stretch)
There’s definitely some weird exceptions like that. From what I’ve seen most stationery, historic planes currently mapped in OpenStreetMap are biplanes from WW2 that are in aviation museums, which aren’t made for long-term habitation. Even in the case 727’s that’s not really their purpose. Whereas at least with trains people often used them as semi-permanent living quarters while they were operational and some were built explicitly for that purpose. Like there’s private rail cars that people have built just to travel around and live in, most of which are no longer in use. They kind of do that with private jets, but then they are still being used and therefore aren’t stationery.
My POV is that mapping ships, aircraft and vehicles as buildings is in most cases a bit of a stretch and can even be considered lying to the renderer. Interestingly, that page has recently been renamed, but I digress.
historic=ship/vehicle etc. is equally awkward, as it’s not a physicial property but instead marks a ship/vehicle as having an undefined quantity of historic value.
I would be in favour of a new main key for vehicles, supported by the manufacturer and model tags that have just been approved. The building=* tag can then optionally be applied to it if the vehicle can be considered to be a building, like that Boeing 727 that Casey_boy mentioned. The same would then go for the historic=* key for historic vehicles/ships like the HMS Belfast.
I could see some reason in having a specific main tag for vehicles, when you cannot enter them. E.g. battle tanks that remained as memorials. Generally, most vehicles likes ships or buses, or airplanes or railway cars, when they are mapped in OSM, are just the same as buildings, i.e. repurposed to fulfill the function of buildings (e.g. apartments, museums/parts). Even if we had a “vehicle” tag, it should be well defined what it should apply to. Maybe it will be more suitable to add “was:vehicle” for many instances
A decomissioned vehicle is still a vehicle, so using the lifecycle prefix for for stationary vehicles would not have my preference. Also, not all stationary vehicles are decommisioned, and it would be practical to have one main key for both decomissioned and not-decomissioned (for lack of a better word) stationary vehicles.
You know, I just noticed there’s upwards of 16 uses of building=houseboat for obviously mapping houseboats. Which just seems weird to me even though I support such tagging more broadly. I think a line should be drawn somewhere, maybe at personal watercraft even if it’s stationery or not.
Like I don’t think someone’s private fishing boat that sit’s at the dock most of the time should be mapped as a building just because it has a small cabin area below the main deck that the owner sleep in sometimes. Same goes for random houseboats. I don’t know how people could know they are stationery anyway. At least with the Delta King in Sacramento it has an address and website associated with it. I assume the same goes for Dieterdreist’s example. Maybe that’s where the line should be drawn, but I don’t think mapping every random object with walls (which some houseboats don’t even have BTW) is at all helpful.
Slightly related, but it’s come up during discussions on the lifeboat-station proposal that amenity=lifeboat has been used ~250 times (apparently in an undiscussed import from an unlicensed source), to map the location of rescue lifeboats when they’re moored at their base e.g. OpenStreetMap. However, it doesn’t always work: OpenStreetMap
e.g. 
