Page MenuHomePhabricator

Consider making a Phabricator project for Wikidata monolingual language code requests
Closed, InvalidPublic

Description

Would it be possible to have a subproject of Wikidata here on Phabricator to keep track of the monolingual language code requests, instead of using the tracking task T144272? I think that would make it easier to keep track of all the requests.

Event Timeline

jhsoby created this task.Aug 8 2018, 1:10 PM
Restricted Application added a subscriber: Aklapper. · View Herald TranscriptAug 8 2018, 1:10 PM

Personally I think we should abolish this Wikidata-specific setting. We should just allow all languages in ULS (which will be the same as in https://github.com/wikimedia/language-data/blob/master/data/langdb.yaml) to be used in Wikidata (as monolingual language, label language, lemma language); wmgExtraLanguageNames (in InitialiseSettings.php) should be abolished too.

All "new language code" request should be focused on adding them to language-data

jhsoby closed this task as Invalid.Aug 8 2018, 1:51 PM

I actually totally agree with that. The system needs a rethinking and reworking, so we should start a task for that instead (if there isn't one already?).

Note "https://www.wikidata.org/wiki/Wikidata:Requests_for_comment/How_to_make_new_languages_enabled_on_Wikidata" says "to add a new label language, add it to MediaWiki"; this is because of they have negative effects. However I doubt how much negative effects it have (why not kill wmgExtraLanguageNames and just use the language-data as the list of available languages). TranslateWiki explicitly don't want to support any ancient language (e.g. Avestan) but Wikidata may like it (especially lemma language is using the same set of label language plus optional -x-Qid suffix, but we clearly should use the ISO 639 code if exist, not mis-x-Qxxx).

In addition we should forwardly review all ISO 639 code (in batch). It is unmanageable if we need to create 7000 tasks to add every languages to Wikidata or a similar database.