Edit language list, remove dialect specificity #848
Conversation
|
Explanation of choices regarding keeping or eliminating dialect specificity: removed most, except in case of Chinese (we already support two dialects as separate locales) and cases where the options differentiate between latin and non-latin characters (e.g., 'iu-latn-ca' and 'iu-cans-ca' for Inuktitut; also Bosnian). |
There was a problem hiding this comment.
PR Review
This PR updates the the list of languages supported by the Zooniverse Translations website.
- Languages are now standardised so that, uh, (trying to see what's the most accurate way to phrase this) only the "main parent class" (e.g. English) is used instead of a specific "dialect child class" (e.g. British English), where possible.
- Exceptions continue to exist where it makes sense, e.g. there's no "main" Chinese language (zh), but there are the Chinese simplified (zh-cn) and Chinese traditional (zh-tw) dialects
Code changes look good, so dev-wise this PR looks fine. Context-wise, I don't see any glaring issues, given my limited understanding of global languages.
As mentioned, follow up is required in 1. PFE (list of languages need to match) and 2. the database (existing languages need to be set to the new standard, e.g. any existing "ne-np" translations need to be "ne" translations.)
| { value: 'ts', label: 'Xitsonga'}, | ||
| { value: 'ts', label: 'Tsonga'}, |
There was a problem hiding this comment.
Fascinating, there are two possible labels for the Tsonga language. Both are valid, though I'm noting that PFE (as of PR 7134) only uses the 'Xitsonga' label
https://en.wikipedia.org/wiki/Tsonga_language
No changes necessary on either this PR or the PF PR, but worth a follow up discussion.
There was a problem hiding this comment.
Duplicate entries allow for more flexible language name search. As implemented in #666, these multiple entries allow users to match via endonym and exonym language names.
For additional related discussion, see: zooniverse/Panoptes-Front-End#6729 (comment), zooniverse/Panoptes-Front-End#6732 (comment)
Note: this multiple label flexibility is a Pandora-only feature, so a single value/label combination must be chosen for the PFE language list. To choose single labels, we defer to the option used or preferred by local speakers.
| { value: 've', label: 'Tshivenḓa' }, | ||
| { value: 've', label: 'Venḓa' }, |
There was a problem hiding this comment.
Similar to my comments on the Tsonga/Xitsonga labels, there are two possible labels for the Venda language. PFE as of the matching PR 7134, only uses the "Xitsonga" label.
https://en.wikipedia.org/wiki/Venda_language
Again, no changes for this PR or the PFE one, but worth a follow up discussion.
There was a problem hiding this comment.
Same logic described here applies in this case.
Towards #845
Current Status: first pass is complete on de-dialect work. Still need to sync with PFE list and remediate any differences. Also: before merge and deploy, cleanup of current translation database entries (i.e., re-assign with-dialect language codes to non-dialect language codes).