|
Hi I am contributing to OSM since years but I still have not found proper way to define name of an area. With area I mean an unsettled (or settled) area which is not an administrative object, nor city nor village. It is just an area in the landscape which local people gave a name. There is lot of such places in Switzerland, like Alp da Munt in Val Mustair. Sometimes I was using place=locality with the name, but this should be used only for single nodes and not for areas. This is important from the rendering point of view. A possibility would be to set area=yes to place=locality, but this is not a recommended solution. Any ideas are very welcome. Milos |
|
Place=locality is quite correct for this usage, named places on earth that do not necessarily correspond to inhabited areas. There's also nothing wrong with that tag being used on an area. In fact it's the opposite: although the current stats show that it is much more frequently used on a node, that kind of data naturally applies to an area, not a node. All those locality nodes are certainly approximations (and often imports). As an example of place=locality usage, that's how townlands are mapped in Ireland. If your favorite rendering displays locality names for nodes but not for areas, you should probably file a bug with them. Don't tag for the renderer. For the default "mapink" style on osm.org, see osm-carto. Note that lately they did a fair amount of area vs node discrepency fixes. Adding "area=yes" for the renderer is not the solution. But you may add a "landuse=*" tag which may help in your case.
(07 Apr '14, 09:54)
Pieren
2
It depends on the feature of course, but I find it unlikely that the whole place=locality deserves a single landuse=* (I expect multiple landuse=* inside a place=locality).
(07 Apr '14, 10:24)
Vincent de P... ♦
I don't think that adding "area=yes" to something that IS an area could be accused of tagging for the renderer, could it?
(07 Apr '14, 11:33)
SomeoneElse ♦
Adding area=yes to a closed way that isn't implicitly expected to be linear (such as one with a highway=*) is as harmless (data-wise) as oneway=no or layer=0, but it is not recommended. If the renderer behaves differently with area=yes on a place=locality closed way, it's a renderer bug. And adding an otherwise useless tag to work around a renderer bug is "tagging for the renderer", even if the tag is harmless.
(08 Apr '14, 11:26)
Vincent de P... ♦
|
