Duplicated Translations

Discussion of Yunit project's internationalization and localization
User avatar
Posts: 124
Joined: Fri Apr 28, 2017 4:00 am

Duplicated Translations

Postby jsalatas » Sun Jul 16, 2017 8:13 pm

As discussed with some translators in telegram yesterday, it seems that for some languages there are duplicate translations in transifex. I'm talking about cases like the ru and ru_RU or es and es_ES translations which as translators informed me shouldn't have any differences. In such cases we should always prefer the generic one (ie ru over ru_RU, es over es_ES) and have specific translations only for cases that they differentiate from the generic (AFAIK such case is the pt compared to pt_BR).

Unfortunately, currently there seem to be some contradicting translations in languages wich shouldn't be (ie contradictions between ru and ru_RU) which I cannot resolve automatically (as I don't speak these languages) and need to be handled by their translators. In particular, translators should check all projects in transifex for dupolicated translations and

1) if it is just the definition but there are no translations made (eg there are many projects that have ru_RU defineds but there are no translations for it), inform me to remove the unneeded language to avoid any further confussion in the future.
2) If both languages have translations, you need to manually check and merge these. It seems that the easier way to do so is to download both translations as files (transifex gives you the ability to do so) and manually compare and merge this. After you have the corrected (merged) file you can then upload it in transifex (it will override the old translations with the ones in the file) and then inform me to remove the unneeded language to avoid any further confussion in the future.

Of course, if you have any questions/problems, you can contact me any time.


User avatar
Posts: 1
Joined: Sun Jul 16, 2017 9:15 pm

Re: Duplicated Translations

Postby YougoCodes » Sun Jul 16, 2017 9:27 pm

My €0,02 (copied from telegram discussion)

there can be nl_NL and nl_BE (for the Netherlands and Belgium respectively). Possibly even more (any Frysians in the house? :-) ) Maybe it's best to just be specific and always use the double naming format. That way there is no confusion about which Spanish would be the 'default' spanish.

Theoretically speaking a Belgian could claim he speaks Dutch (nl) as much as a Dutch person, but a Dutch person calls an umbrella 'paraplu' (french borrowed) while a Belgian would say 'regenscherm'. If there's only the nl locale, that word could cause back and forth editing untill someone requests nl_BE to be created. My thought was to skip that and be clear off the bat which nl is available.

John said:

Actually my point is that if you decide that there should be nl_NL and nl_BE then there shouldn't be a generic nl. And someone needs to check if there is currently something under nl that needs to ne merged to either nl_NL or nl_BE (or both).

To which i said:

I agree, but maybe take it one step further and say there should never be a generic nl (or es/en/ru/...) as there could rise disagreement on what generic is supposed to be.

Return to “Translations and Localization”

Who is online

Users browsing this forum: No registered users and 1 guest