Weblate is an open source project to help software developers translate their projects. Translators can work in a web interface, and not have to install third-party software or use git or GitHub.
I have been working on integrating the plone.app.locales package in Weblate, and now all translations contributed through Weblate are pushed to a given branch of this package, which I regularly monitor and merge into the main package.
I want also to do the same with Volto, but I have some issues to fix before announcing it as an official way to contribute, and I want to fix them as soon as possible.
I tried this out while polishing up the docs for publication. I would love feedback about the docs as well. I tried to make the process clear, especially for first-time contributors, because almost all of us are first-timers to Weblate.
It's a nice way to spend a few spare minutes while waiting for buildout or acceptance tests to run. Once I understood the process, and how simple and elegant Weblate made it, I wanted to keep going.
Q: why don't we use AI for translating the missing strings, generating a proposal for a human reviewer?..this would be much faster and likely more accurate than translating the missing string manually or getting them translated at all manually. The accuracy of translation made by e.g OpenAI or DeepL are in general (for the most common languages) much better than human translations. Also, the days of nonsense translations where translations systems had no context about the problem domains, are gone.
I did a quick test using LibreTranslate and translating the missing strings into French:
And this is the automatic translation into German:
Perhaps I would need to adjust my script to properly handle source strings, because sometimes we have the source string in the msgid and sometimes in the default comment, but this works as a proof of concept.