Translation services such as Apertium provide a good start for users to create their translation. Since users are correcting those, it would be great if those corrections could be easily accessible so that teams developing the translation services could improve them based on that feedback.
As part of this task a discussion on the information to capture and how to expose it is needed. Some fields that may be helpful:
- Source text
- MT translation proposed
- Final translation by the user
- MT provider
- Editing distance (so that we can filter big/small differences)?