Based on the following section:
Should I only add provenance information to each documentation page, or should I also include it in the README.md file of the source repository?
You should do both, for maximum visibility.
1 Like
Regarding manually maintained translation vs. automated translation:
If readers intend to read a document translated entirely by machine or AI, they can already do so using their browser’s built-in translation tools, such as Google Translate, or third-party extensions like Immersive Translate.
However, I’ve always believed that no matter how much translation tools improve, the final decision on how to translate something should always be made by a human during proofreading. The goal of the l10n project I’m working on is to provide a platform that makes it easy for users to perform manually maintained translation.
BTW, Crowdin actually provides a feature called “Pre-Translation via AI”. See the following links for more details:
Currently, for all l10n projects, I’m using the most basic Crowdsourcing workflow:
Since using AI-powered features requires a paid subscription, which is not included in the Crowdin Open-Source License, I don’t plan to look into this feature for now. If I secure enough funding in the future, for example through sponsorships, I believe this feature could significantly accelerate the translation speed and efficiency.
All previously mentioned issues should have been discussed by now. If someone has any further questions about the ros2-docs-l10n project, feel free to ask.