Say the supermarket downsizes and the other half of the building becomes used by another company. We split the building, creating 1 new building, glued to the original building. Now there is a 50% chance the existing building outline continues as the supermarket and 50% it's the new company continuing life in OSM with that way id.
I'm working on our local bus stops a lot and they all have a ref displayed on them, which is very convenient to refer back to them. In the other side of the country and in Brussels, we have to make do with the name. If we ever get access to the underlying DB, we might be able to add the internal refs, which to me is still the most convenient way to identify specific bus stops. Right now there are usually at least 2 bus stops with the same name on both sides of the road. relation membership could make them unique, once route relations are added, but it complicates matters quite a bit to have to identify them that way. (I'm not talking about linking bus stops to Wikidata/Wikipedia). I need to link back to them to compare with the data coming from upstream. Or to assist in creating/verifying route relations. All that to say that, to me, ref tags on our objects seem like the most stable and practical way to identify them. Polyglot 2013/5/7 Peter Wendorff <wendo...@uni-paderborn.de> > Unfortunately I'm too busy to investigate how much elements in OSM > change their meaning instead of deleting the old and creating the new > object. > > In addition your "fooling the concept" is not correct. If a supermarket > is abandoned, and was tagged as a building + name + shop=supermarket, > you would not delete the whole object and add a new object with the > building tag alone, right? > > regards > Peter > > Am 07.05.2013 10:25, schrieb Stefan Keller: > > 2013/5/7 Peter Wendorff <wendo...@uni-paderborn.de>: > >> Look what happens in OSM all the time: POIs are moved slightly to match > >> aerial images - following your definition that should be another ID now > > > > No, That's one of the nice properties of ids without coordinates! > > To me it would remain the same - except when a tool or the user is > > fooling the concept. > > At least the tools you can debug. > > > > Yours, Stefan > > > > 2013/5/7 Peter Wendorff <wendo...@uni-paderborn.de>: > >> Am 07.05.2013 09:58, schrieb Stefan Keller: > >>> Hi, > >>> > >>> You wrote: > >>>> - it's roughly in that bounding box (e.g. the city or a given part of > >>> > >>> A soon as you use the word "roughly" - the id approach is doomed to > fail. > >>> According to OO and database technology an id is a well-defined > >>> surrogate with a well-defined data type. > >> > >> Then it's not the same "permanent" we talk about. > >> Look what happens in OSM all the time: POIs are moved slightly to match > >> aerial images - following your definition that should be another ID now > >> - but that's not what people usually want if they request for a > >> permanent ID, similar to changes from node to polygon to multipolyogn > etc. > >> > >> "It's in that bounding box" nevertheless would have been the better > >> wording, (equalling "is roughly at that position, so if you want to use > >> roughly/estimation, it's possible even then). > >> > >> regards > >> Peter > > > > > _______________________________________________ > talk mailing list > talk@openstreetmap.org > http://lists.openstreetmap.org/listinfo/talk >
_______________________________________________ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk