We will need a definition of "building". Some may consider a terrace of houses to be a single building.
One definition I have worked with involves assessing the ability of the "building" to remain standing and usable if the "buildings" on either side were removed. If a house in the middle of a terrace was destroyed or demolished, would the adjoining houses still be structurally sound? If so, they are separate buildings. If not, the terrace is one building. This also has meaning in the vertical dimension. In a block of flats, the units cannot be called buildings because they are structurally dependent on each other. Having said all that, using a single outline for a terrace of houses is an adequate first-order approximation for many purposes. Further refinement by splitting into individual dwellings is a possible "next step" if the information is available and if a mapper has time. --colin On 2016-03-16 17:12, Blake Girardot wrote: > I am reluctant to suggest that mapping large groups of buildings as one > outline is a good idea. As I said, to me it is a last resort and should be > avoided at all costs. Otherwise we are going to get blocks of easily mapped > buildings outlined as building just because that is a lot easier and then > leave the detailed mapping to someone else. > > But that being said, since it is acceptable to map large blocks of buildings > as building in some circumstances, I think we might need a new value to the > building=* key. > > Most data consumers that I work with usually consider building=yes to > indicate one building. > > If we know we are mapping multiple buildings under one polygon would it be a > good idea to add something like a 'multiple' value to the key? > > building=multiple > > That would also make it easier to locate them for further refined mapping in > the future. > > cheers > blake > > On 3/16/2016 4:48 PM, althio wrote: Simon Poole wrote: IMHO we always allow > and support progression from rough to more detailed. > +1 > > Philip Barnes wrote: Mike Thompson wrote: My feeling is that individual > buildings should be mapped. > > In an ideal world I would agree, but we don't live in one and in some cases > such as medieval building layout it can be incredibly difficult to work out > what roofline belongs to which building. > > I would say its ok, and better than not mapping buildings at all, then you > can always improve it after more surveys. +1 I agree it is good to have rough mapping and let it improve over time. Back to OP question: Once you trace a rough outline for multiple buildings, what is the tag? building=yes? or another value? with a note=*? a fixme=*? - althio _______________________________________________ Tagging mailing list Tagging@openstreetmap.org https://lists.openstreetmap.org/listinfo/tagging _______________________________________________ Tagging mailing list Tagging@openstreetmap.org https://lists.openstreetmap.org/listinfo/tagging
_______________________________________________ Tagging mailing list Tagging@openstreetmap.org https://lists.openstreetmap.org/listinfo/tagging
