Page MenuHomePhabricator

[GOAL] Improve experience around adding new language codes for Wikidata
Open, Needs TriagePublic

Description

work in progress

User story:
As a Wikidata architect interested in adding a new language code,
I want a process that lets me know in advance if a new language code will be accepted or not
to not waste time.

As a Wikidata developer,
I want a more failproof process
to not make errors (see e.g. T273627).

Problem:
Our current processes around adding new language codes sometimes lead to inconsistent and hard-to-predict results. Also, the implementation can be inconsistent in some cases. This can be a frustrating experience for all parties involved and lead to subpar results.

Also: we have seem to have some bad config T284808: Add a configuration variable that allows disabling language codes for labels, descriptions, and aliases

Examples:

Solution:
Make processes clearer and easier. Ideally one could predict the outcome of a request even before sending it in.

Also, we need a process that helps to vet the changes to avoid back-and-forth decisions or problematic outcomes (e.g. removing code with existing Lexemes)

Related:

Event Timeline

Manuel renamed this task from Improve Community experience around adding new language codes for Wikidata to [EPIC] Improve Community experience around adding new language codes for Wikidata .Jul 28 2022, 7:08 PM
Manuel added a project: Epic.
Manuel renamed this task from [EPIC] Improve Community experience around adding new language codes for Wikidata to [GOAL] Improve Community experience around adding new language codes for Wikidata .Jul 28 2022, 7:12 PM
Manuel removed a project: Epic.
Manuel renamed this task from [GOAL] Improve Community experience around adding new language codes for Wikidata to [GOAL] Improve experience around adding new language codes for Wikidata .Aug 3 2022, 1:02 PM
Manuel updated the task description. (Show Details)

After T341409: [TECH] Use LanguageNameUtils::ALL for monolingual text and lexemes has been implemented, this should be slightly easier now: Lexemes and monolingual texts now share their supported language codes.

Very cool! Thank you for the update @hoo!