For some countries it will be useful to filter also by regions inside the country

Since when CLDR provides such data?

In T133029#2221735, @Nemo_bis wrote:

For some countries it will be useful to filter also by regions inside the country

Since when CLDR provides such data?

I didn't mention CLDR :)

I don't know how exactly we shall do it. It requires two things:

• Fine-grained location. We may have it in the Wikimedia geolocation service, but I haven't checked deeply.
• Mapping between country sub-divisions and languages. Maybe we can just do it manually for countries that have more than ten languages, with some help from Wikipedia, Ethnologue and common sense.

Note that some countries have several common languages, and that some countries even can have differences among closely related municipalities. In Norway Norwegian language is the most common language, but in Kautokeino municipality Northern Sami language is a common (possibly the most common) language.

We could simply make a language raster of the globe and adapt the language list to whatever are most common in each subarea. This has the added benefit of relaxing the requirement to follow official languages. Some users would probably argue that official languages should be used because they do not include specific languages.

Note similarity with T140453: Prioritize shown languages based on article's geographic coordinates

The requirements must be clarified. I vague recall we discussed something about the scope in the CLL document and that the tickets was supposed to be updated to reflect that.

Do we need to check what the cookies provide? Do we need to improve those systems?
Do we need to collect languages per location? For all world?
Do we need to implement finding languages per location? For PHP? for JavaScript?
Where should we use that information?