[Tagging] wikipedia links and copy + paste in tag definitions

[Martin Koppenhoefer]

2017-05-01 14:43 GMT+02:00 Andy Mabbett <andy at pigsonthewing.org.uk>:

>
> > no, my main concern is that both have different scope, and that people
> are
> > modifying wikidata without looking at the use of osm tags.
>
> You wrote "sometimes we have multiple tags in OSM each covering a
> specific aspect or subtype of what is covered by a single WP article".
> I think most reasonable people would regard that as complaining about
> a lack of complexity
>



First thing I wrote in the first post was about having different scope. The
part that you cite is not necessarily about lack of complexity but rather
about different organisation of information.



>
> I asked you for examples of damage caused by Wikidata definitions
> changing; I note that you have not provided any.
>


you don't need any example to understand that definitions of OSM tags
changing outside of OSM and OSM context will very likely not be what we
want.



>
> >>> You have to follow a lot of links and definitions and there
> >>> are many dependencies, everything is connected:
>
> >> It's called "linked data" for a reason.
>
> > yes, but it makes it completely impractical for osm users to use it as
> > definitions for their tags.
>
> It does no such thing; as evidenced by the widespread and growing use
> of Wikidata IDs in tag infoboxes, about which you complained.
>


I complained because I noted this growing use which leads to
inconsistencies and might in the long run invalidate our established
definitions if people start to tag accordingly ("I used this tag because
there was a wikidata link which made me believe the tag was for this kind
of object") and after a while we will find that "this tag is not useful any
more because a lot of different things are tagged with this same tag, there
is no consistent use". This is something that already happened in the past,
it is important that we give consistent advice in the wiki to reduce this
kind of data invalidation.



> >>> E.g. place in osm is orthogonal to administrative entities, in wikidata
> >>> it is not. Add a property/"instance of" to the wikidata town object
> like
> >>> "administrative territorial entity" and you changed all towns.
>
> >> If someone made such a ridiculous change, they would be reverted. This
> is
> >> FUD.
>
> > The entity "town" ALREADY IS an administrative territorial entity in
> > wikidata, and nobody has reverted it: https://m.wikidata.org/wiki/Q3957
> > At least since 2014, and a lot of people have "automatically patrolled"
> it
> > since then, whatever that means,
>
> So, not a sudden and after-the-fact change, which is what we were
> discussing, then?
>


back then it was a sudden change, because this wasn't in before. It doesn't
matter whether this change happens suddenly "just now", or at some time in
the past or future. The fact that it wasn't reverted "proves" that your
previous assumption about timely revertion was wrong, which highlights the
very real risk, no?

Cheers,
Martin
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/tagging/attachments/20170501/39473686/attachment.html>