[Tagging] part_of:wikidata key

Kevin Kenny kevin.b.kenny+osm at gmail.com
Tue Nov 28 23:14:41 UTC 2017


On Tue, Nov 28, 2017 at 5:29 PM, Christoph Hormann <osm at imagico.de> wrote:
> The only advantage i see is from the wikidata side that the
> responsibility of dealing with a data change is superficially delegated
> to the mapper in OSM with wikidata tags but it is an illusion that this
> will actually lead to reliable ID maintainance since ultimately this is
> something you cannot require the mapper to do because it is
> non-verifiable information.  OSM recruits mappers to map the verifiable
> world, we do not require them to also research secondary source
> references in wikidata to make sure they correctly maintain wikidata
> IDs.

Well, I certainly haven't tried to maintain them, but I encounter them
on objects. That's usually because I've worked with the boundaries
of a park, add a Wikipedia link, and find that Wikidata gnomes come
after me and add a Wikidata link with it..

If the gnomes are going to add wikidata=Q314159265 to some
object. I'll happily give them an Overpass query to retrieve that
object (any any object tagged alike). If the gnomes are going to
ask me to map things that I can't see in the field (or do a quick
search for: I do like to link to parks' Web pages and Wikipedia
articles), that becomes unreasonable.

If someone else wants to maintain data on objects that I mapped,
fine. I'll happily give them back what they gave me. If someone expects
me to add facts that I don't know, according to a schema that I
don't understand, sorry, I won't do that.

Since I can't give someone a stable key to look up an object that
they want to remember, the best I can do is let them hang their
own key on it, and give it back to them. You think it's worthless because
it isn't perfect.  I think it's better than nothing.



More information about the Tagging mailing list