[OSM-talk] "Keulen" (was Re: Can wikidata links help fight name inflation?)

Christoph Hormann chris_hormann at gmx.de
Sat May 30 10:00:06 UTC 2015


On Saturday 30 May 2015, Frederik Ramm wrote:
>
> Whoever does the preprocessing is of course free to join OSM data
> with any other database; so if for example we agreed that Wikidata
> was our name repository of choice and we'd reduce our own name:xx
> tagging to what is used by the people living in a place, then there
> would have been a wikidata tag on the "Köln" node and the software
> that generated your offline database could have made a choice
> whether, and to what degree, to include the Wikidata name catalogue
> in your offline data set.

I am all for going strictly by the verifiability principle but this 
should not require local verifiability.  There are many fields where 
this does not work at all:

- maritime features outside territorial waters.
- areas without local population or where local population has little 
contact with the outside world.
- touristic areas with little local population like remote high mountain 
areas.

In all of these cases names of features will be very difficult or 
impossible to verify locally in any language, verifiable names here 
primarily exist in written records abroad and in the memory of people 
who have visited the area but who are definitely not locals.

And IMO it would be very regrettable if OSM universally referred to an 
external database for names of such features and specifically excluded 
them from the scope of the project.  The question is also would you 
leave such features without any name and if not on what basis do you 
decide which name to tag?

My suggestion for the whole issue would be to clarify the verifiability 
rule w.r.t. names and advise mappers to document sources of non-local 
names they add in changeset comments and source tags and ultimately in 
cases of conflicts it would be the burden of proof for the mapper to 
show the names added are verifiable (like by being consciously and 
specifically used by several other independent sources).

-- 
Christoph Hormann
http://www.imagico.de/



More information about the talk mailing list