Open cbjeukendrup opened 1 year ago
Yes, I'd remove Arabic (Algeria) (ar_DZ)
and I guess Arabic (Egypt) (ar_EG)
too.
Propbably some others too, like Igbo (ig)
, Kabyle (kab)
and Sicilian (scn)
, but those at least don't harm.
Only exception from that 60% rule should be English (US) (en_US)
We could also 'seed' (or merge) Arabic (Algeria) (ar_DZ)
and Arabic (Egypt) (ar_EG)
with Arabic (ar)
and so get them to a reasonable state right away.
Also I think we should change Mongolian (Mongolia) (mn_MN)
to Mongolian (mn)
and to Hindi (hi)
Yes, I'd remove
Arabic (Algeria) (ar_DZ)
and I guessArabic (Egypt) (ar_EG)
too. Propbably some others too, likeIgbo (ig)
,Kabyle (kab)
andSicilian (scn)
, but those at least don't harm.
I see Sicilian, interesting. I know this thread is a year old, but, just for curiosity, were expected other translations of MuseScore in other Italian dialects?
Actually rather than removing translations we (@cbjeukendrup and I) meanwhile had the idea of having cascades of fallbacks rather than just falling back to English. For example the lat Sicilian falling back to Italian first before finally falling back to English. Same would then apply to other Italian variants. And those Arab national variants to plain Arab first, then to English.
See #23638
The problem Currently, translations for 68 languages are packaged with MuseScore. However, only a fraction of those is being actively translated. At the moment of writing, only 13 languages are 100% translated, and 16 others have less than 2000 untranslated strings (of the 7845 total strings).
It's not only a bit misleading to include "translations" that leave half of the app untranslated; in some cases it's just problematic. For example, we have
Arabic (ar)
, which is very reasonably translated with 65%, but we have alsoArabic (Algeria) (ar_DZ)
, which contains exactly zero translated strings. Users from Algeria will get this variant, with no translations at all, instead of the properly translated "generic" Arabic.This problem also prevents us from accepting new languages: we need to be sure in advance that someone will indeed work dedicatedly on the translations for that new language.
The solution I propose that we introduce a threshold on the translation percentage for languages to be included in releases. Let's choose 60% to start with.
This way, translations that are not really translations are excluded until they get updated, and we can freely accept new languages on Transifex, without worrying whether they will be translated enough.