[Tagging] part_of:wikidata key

[Christoph Hormann]

On Tuesday 28 November 2017, Jo wrote:
>
> Grouping objects in associatedStreet and river relations (and adding
> wikidata identifiers on them) partly solves this, but
> associatedStreet wasn't liked much and is pretty much dead in the
> water and in the river relations, only the linear features are
> included, not the riverbank area ways.
>
> Does it make sense to create pages for this on the wiki? Or would
> several people have issues with such a proposal?

We had a discussion some time ago on the talk list regarding how the 
idea of wikidata IDs already clashes with the idea of on-the-ground 
verifiability and the established rule not to have external IDs in the 
OSM database.  My current view of that discussion is largely outlined 
in:

https://lists.openstreetmap.org/pipermail/talk/2017-October/079126.html

Having additional tags representing abstract concepts and categories of 
features would further exacerbate this problem.

If what you want to tag part_of:wikidata representas a spatially defined 
and as such verifiable concept you should map it in OSM and what 
features are part of this derives from the spatial relationship and 
existing tags in OSM - in other words part_of:wikidata would be kind 
of 'is_in all over again'.  If what you want to tag with 
part_of:wikidata is a class of objects or category you should look for 
an existing tag or invent a new tag describing this class/category and 
you can add the wikidata ID to the tag page on the wiki describing the 
relationship (which again of course is not necessarily a 1:1 
relationship).

Tagging something like part_of:wikidata instead of either of this looks 
to me like delegating the recording of mapping information and with 
that the authority of defining concepts and classes to an external 
authority (wikidata) with ultimately incompatible standards.

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