I’d like to update you on the recent progress that has been made in terms of localization of the Sentence Collector as well as how we handle languages.
For a long time the Sentence Collector had languages enabled that had not been officially requested, and were never on track to be added to Common Voice. This also meant that new languages had to be enabled in both Pontoon and Sentence Collector, adding overhead to the process. Only two languages had > 10 sentences contributed, without being officially added to Common Voice. I created Kannada and Sanskrit for this.
With the latest release only languages officially enabled on Common Voice (done in Pontoon) are also enabled on Sentence Collector. This means that when Hillary adds your language to Pontoon after you requested it, it will also automatically be enabled in Sentence Collector around 30 minutes later. Initially the native language name might not be shown, but it will fall back to the language code until localization and syncs are done.
I hope that with this change we can better streamline the language process.
Link to language addition process
There is now a link to the language addition process next to the dropdown for the contribution language:
UI Language Dropdown
We have added a language selection dropdown for the User Interface to the profile:
For now this will be saved in local storage, but not on the profile itself. This means that if you log in on another device, you will need to set your preferred UI language there as well.
Improvements for localizations
The strings now have “group comments” to better identify where in the Sentence Collector they are used. I also added a note for the Keyboard Review Shortcuts to be clear that the shortcuts itself are not localizeable yet.
Deployments of localizations to staging
We are now deploying to https://commonvoice.allizom.org/sentence-collector/ 4 times a day. Each deployment will update the strings from Pontoon. This allows localizers to verify their changes in Pontoon and do quality control. Thanks to @rprys for pointing out a bug with that!
Note that this is a staging environment, these sentences will NOT be exported to Common Voice. Only use this for testing purposes!
Validation error messages
The default validator error messages for the “Add” process can now be localized as well.
Note for developers of the Sentence Collector
Please go through the Localization Documentation as it provides information on how to add and change strings.
Happy to hear your thoughts!