Frederik Ramm wrote: > >> 2. Treat relation membership as a characteristic of the members. > > I do not think this is a good idea. It seems conceptually wrong to me. > If the public transit authority creates a new bus route, does that > really change all the roads along which the bus runs? Is the road any > different on the ground the day after the bus route has been introduced? >
I believe the answer is somewhere in between. It is true that adding it to a relation is not a change to the core data of a way. But having an indication in the way that relations exist and which ones they are would be a tremendous improvement for handling and understanding the data. Especially if you are not a power mapper and are not even aware that the object you are handling may be involved in relations or if you do not have a local PostGIS database or a complex editor to already do the searching for you. Maybe the way to go is an optional part with metainfo about relation membership which can be requested with an additional parameter or alternate call. The API should have that information readily available when it compounds the data. In any case, such double linking information should be ignored when an object is uploaded to the API to avoid conflicts between way->relation and relation->way linking. bye Nop -- View this message in context: http://gis.638310.n2.nabble.com/A-proposal-to-improve-relation-handling-tp6714787p6719356.html Sent from the General Discussion mailing list archive at Nabble.com. _______________________________________________ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk