As a side question to this : why do we even need to import languages from Wikidata ?

.I don't feel comfortable with a massive import mainly because among all the elements that are languoids, there is a mix between language families, dialects, varieties, etc. And so I wonder about our means to manage all of this. So agree to import list of "official" langages but skeptical about a massive import

Clearly importing *all* languages (lato sensu) from Wikidata doesn't seems to be a good idea. But on the other hand, it would make sense to import some "important" missing languages. Between 600 and 10 000, there is probably a sensible middle-ground.

The important distinction here is: what "important" languages do LinguaLibre miss? What are the ones who are most likely to get records?

PS: with a SPARQL query, I found that there is 595 languages right now in LinguaLibre with an item, but only 164 lang have at least one record (including 68 languages with less than 10 records), so 431 languages have an unused and currently-useless items. Not sure how much not having an item is really a blocking code, do we have actual feedbacks.

In T284645#7145645, @Poslovitch wrote:

As a side question to this : why do we even need to import languages from Wikidata ?

What solution do you propose instead?

In T284645#7149939, @Pamputt wrote:

[...] So agree to import list of "official" langages but skeptical about a massive import [...]

I agree, there are unwanted elements among the 10k elements, we have to find a middle-ground as said by Vigneron.
Here is a list of living languages with an iso 639-3 code on Wikidata : https://w.wiki/3Ugk (Edit: thanks @VIGNERON for the correction 😉). This is probably not precise enough, but it can still be a beginning, what do you think of it?

In T284645#7153199, @VIGNERON wrote:

[...] PS: with a SPARQL query, I found that there is 595 languages right now in LinguaLibre with an item, but only 164 lang have at least one record (including 68 languages with less than 10 records), so 431 languages have an unused and currently-useless items. Not sure how much not having an item is really a blocking code, do we have actual feedbacks.

Every language deserves to be recorded on Lingua Libre, and IMO users should not have to ask admins to add a language, the language should already be there, or at least they should be able to add it themselves. I find very inconvenient in terms of user experience to have to ask to humans (which can take up to a few days) to add one's language.

Isn’t it possible to only use the Qid from wikidata, like for locations?