Page MenuHomePhabricator

Detect unedited content copied out of Content Translation
Open, MediumPublic


Based on input from Vietnamese editors, some users avoid the Content Translation restrictions to publish unedited content by copying the contents out of the tool into a new article. This could lead to low quality translations to be published without any easy way to filter them for easy review (since they will be tagged as any other article created from scratch).

The original request was to disable copying form Content translation, but that could negatively affect the translation workflow of those using the tool properly. Thus, we want to explore ways in which this problematic pattern can be detected.

This ticket proposes to include a special tag as part of the HTML content copied from Content Translation. The tag will be added only when the content of the translation has a high percentage of unmodified MT (based on the same thresholds used for showing a warning to the user). In this way, when the content is pasted in a wiki, edit filter can be used to react to it. For example, adding it to a category for easy review, or block the publication of it.

Event Timeline

Pginer-WMF triaged this task as Medium priority.Mar 24 2021, 9:56 AM
Pginer-WMF created this task.

I've provided another sound alternative solution that is easier to implement (technical aspect). Pginer and I had a discussion about the pro's and con's of each solution. See here.