Prune transifex project

[StanFromIreland]

There are quite a lot of languages on transifex, quite a few should be removed because:

1. There is no corresponding translation repository and the transifex translators are not active
2. The translation no longer uses transifex
3. Language is not supported by Sphinx

Crossed of languages stay, those not crossed off will be removed with due process. Translation coordinators will be contacted to see if they are still using transifex.

• Russian -- 87%
• Chinese (China) -- 82% PENDING REMOVAL
  • Transifex python/python-docs-zh-tw#1031
• Portuguese (Brazil) -- 61%
• Ukrainian -- 54%
• Japanese -- 54%
• Polish -- 11%
• Spanish (Spain) -- 10%
• Indonesian -- 9%
• Persian -- 6%
• Chinese -- 3% PENDING REMOVAL
• Korean -- 2% PENDING REMOVAL
  • Transifex translation python/python-docs-ko#1073
• Chinese (Taiwan) -- 2% PENDING REMOVAL
• Czech -- 1%
• Arabic -- 1%
• German -- 1%
• Hungarian -- 1%
• Spanish -- 1%
• Turkish -- 1%

0% Translation Projects

• Vietnamese -- 0%
• Chinese (Hong Kong) -- 0%
• Spanish (Argentina) -- 0%
• Kyrgyz -- 0%
• Pashto -- 0%
• Spanish (Colombia) -- 0%
• Italian (Italy) -- 0%
• Albanian (Albania) -- 0%
• Hindi (India) -- 0%
• Indonesian (Indonesia) -- 0%
• Korean (Korea) -- 0%
• Marathi (India) -- 0%
• Portuguese (Portugal) -- 0%
• Sinhala (Sri Lanka) -- 0%
• Spanish (Latin America) -- 0%
• Central Atlas Tamazight -- 0%
• Chinese (Mandarin) -- 0%
• Arabic (Egypt) -- 0%
• Chinese Simplified -- 0%
• Chinese (Singapore) -- 0%
• Arabic (Iraq) -- 0%
• Bengali -- 0%
• Bengali (India) -- 0%
• Catalan -- 0%
• Danish -- 0%
• Greek -- 0%
• English (United Kingdom) -- 0% PENDING REMOVAL
• Finnish -- 0%
• Hindi -- 0%
• Italian -- 0%
• Georgian -- 0%
• Lithuanian -- 0%
• Norwegian Bokmål -- 0%
• Nepali -- 0%
• Dutch -- 0%
• Portuguese -- 0%
• Romanian -- 0%
• Urdu -- 0%

[rffontenelle]

One more reason to prune: language variant when both variant and fallback language are low activity, and the split effort doesn't justify. (E.g. Portuguese and Portuguese from Portugal)

[StanFromIreland]

There are several we should remove because of this e.g. English (United Kingdom)

[AA-Turner]

en-GB is often useful to test translation infrastructure for English-speakers. No massive opposition to removing it, but I'd argue if we want to do localisation properly we should be supporting regional variants of language where possible.

In the large, I think (1) is a weak rationale for removing a language. Having a language in transifex doesn't hurt us, and it's easier for a prospective translator to get started on a pre-existing, but 0%/1% language, than it is to request everything be set up from scratch.

If a translation doesn't use transifex, that makes sense to remove. If a language isn't supported by Sphinx, that just means that slightly more work is needed (adding the translations to Sphinx), and shouldn't be a hard blocker.

A

[StanFromIreland]

en-GB is often useful to test translation infrastructure for English-speakers

How is it used for testing if it is not pulled to any repository? It also has no translations. There is nothing against regional variants, we are removing them now because no one wants to translate them, if someone wishes to do so in the future adding it is trivial.

having a language in transifex doesn't hurt us

It may not hurt us, but it hurts translators who waste their time on a project with no repository. The translations are not used since they are not pulled anywhere, and it is difficult to get their attention (Transifex messages can easily be missed)

t's easier for a prospective translator to get started on a pre-existing, but 0%/1% language, than it is to request everything be set up from scratch.

You should probably look at the "Request a Language" button on transifex, it is quite straightforward, but thats beyond the point. This is currently not desirable, as the translation needs a repo. See for example the Chinese project on transifex, translations are added daily but the translations are not used, since the project now uses prs! (The users have missed my messages)

The translator has to set up everything, otherwise it is pointless as they have to hope another translator will do it for them and using their translations -- at the end of the day, someone has to do it. We have been working on building documentation and adding template workflows (WIP, see pr) to make this as easy as possible.

hard blocker.

Maybe this issue was not clear enough, we do not mind adding translations! There are no "critera", we simply want to clean up to ensure translators do not waste their time and efforts.