[OSM-talk] Naming disputes in Ukraine

[Peter Wendorff]

Am 25.07.2012 14:54, schrieb Miloš Komarčević:
> Which brings us back to the point Peteris Krisjanis made:
>
> name= without the context of a language is somewhat useless from the
> database point of view, apart from quick and dirty rendering of what
> is on the ground.
+1
But it is useful for quick and dirty usage like "give me any name" - and 
there it's much better than to use the first name-tag found (hooray, 
everywhere in the world we get Albanian names on the maps!) or something 
like that.
> A much better approach would be to dispose of it, and force having
> name:<lang> everywhere.
I would not like to dispose name, but enforce (but not hardly require) 
to add name:de even in single-language parts of Germany and so on.
Done completely that would lead to
name=* everywhere in any language that seems to be useful for the user 
tagging it (doesn't prevent the disputes of course, but neither does a 
mapper-defined lang-attribute).
name:*=* everywhere as much as needed, but enforced to be once for the 
language used in name. This can be used to determine the language of 
name by comparison, while some times more than one language might match.
Therefore it's redundant, but matches the requirements fulfilled by the 
lang-tag implicitely.
> Then one would be able to define e.g. on the
> administrative level (country, district, municipality) what languages
> to use/render on objects inside that relation.
>
> For example: for the whole of Italy relation you would have "lang=it",
> but for the South Tyrol you would have lang=de;it (or whatever order
> is appropriate) which would take precedence. For any exceptions, you
> would add lang= on the object itself which would have highest
> priority...
-1
I oppose to have rendering rules in the data like suggested here, but 
again the implicit language definition as described above may be used 
here, too - even if probably not in the main maps renderer, but that's a 
completely different thing.

regards
Peter