[OSM-talk] Naming disputes in Ukraine

Aun Yngve Johnsen lists at gimnechiske.org
Wed Jul 25 18:04:03 BST 2012


I do not agree with you on this, besides, the language polygon I mentioned can solve your desire here. If the renderer knows what is the official language in an area, than you can in theory tuggle freely between that/those official languages and any other language you would desire. For instance if I am in a country not using latin alphabet street signs, I still would like to see the names in latin script.

IMO all of this is not up to OSM, but to renderers. A tag in border relations should be enough to indicate official languages. After that it is up to the renderer how to solve it. Doing it this way relieves OSM from any naming disputes (we are thus handling all names equally), and renderers choose how much effort they want to put in name rendering.

Your tourist rendering can easily be done inside an app where official language have been set in a boundary relation. This might also allow municipal relations to overide national standards, as I believe not all parts of Belgium have the same order between the three official languages rendered on street signs. The same can apply for countries that have more official languages only in certain regions, such as northern Norway also have Samii names, the southern part of Finland use a lot of Swedish names, various regions of Switzerland have different settings, etc.

Your approach I feel is exposing OSM to a lot of politics and possible edit wars, while the option I try to suggest might limit this type impact for OSM.

Same also, for all those countries where tthere are only one official language, the language tag in the boundary relation can allow the renderers to assume that name= is the same as name:xx even if name:xx isn't tagged. That way it will not be confused with name:yy and name:zz

Aun Y. Johnsen
Sent from my iPad

On 25. juli 2012, at 13:37, talk-request at openstreetmap.org wrote:

> Message: 3
> Date: Wed, 25 Jul 2012 16:30:47 +0000
> From: Svavar Kjarrval <svavar at kjarrval.is>
> To: talk at openstreetmap.org
> Subject: Re: [OSM-talk] Naming disputes in Ukraine
> Message-ID: <50101F37.300 at kjarrval.is>
> Content-Type: text/plain; charset="iso-8859-1"
> 
> It does fall under OSM's jurisdiction to indicate what the official
> names are and which are translations. If I'm a tourist in a country, I
> need to know the names on the signs, not a renderer's guess or my native
> language's name.
> 
> If the renderers have to guess, they have to create additional data for
> each area and research which languages they should use in each of them,
> instead of focusing on the rendering process itself.
> 
> - Svavar Kjarrval
> 
> On 25/07/12 16:10, Aun Yngve Johnsen wrote:
>> In my opinion sorting languages for rendering is the renderer's problem, one can assume that name= tags in countries with a single language is the national language, but for a renderer to understand this, poligons with lang=* or similar must exist (either within OSM or in a separate database)
>> 
>> It will be much more logic to store every name in name:xx tags, and let renderers sort out how to deal with them. Renderers must thus have fallback rules in places where several language name tags exists, but again, this is the renderers problem.
>> 
>> Now, to allow completely I18N compability in the maps, one would need every version of names to be available, either in name:xx tags within OSM, or in a separate database for name translation. This way, OSM could be completely neutral to naming disputes (as no 'default' name would exist in our database), leaving renderers resolving the various problems. I would love to see our 'default' map layer to show names based on browser language settings (i.e. Moscow would show as Moskva on my map, and my home town show up in Cyrlic in the browser of anyone from Moscov)
>> 
>> I understand that it might be a long and complicated task cleaning this up, as 'the entire world' is tagged with name= and only a few regions and places have aditional name:xx
>> 
>> Aun Y. Johnsen
>> Sent from my iPad



More information about the talk mailing list