Let's chat about if we need to do mocks / or have a UI component to how we want this A/B test to be run.
|Open||None||T118278 [EPIC] Improve Language Identification for use in Cirrus Search|
|Open||None||T140292 A/B Test TextCat settings on non-WP projects|
|Declined||None||T142584 Investigate creating mocks for UI component of this A/B test|
By default we could use the same cross-wiki UI we did for the other language ID A/B tests. It's not great, but it is the same set up, so we'd be comparing apples to apples. OTOH, a better UI might mean more people understand it and get value out of it. I don't have any strong opinions about the UI, though a clearer separation between the local results and cross-wiki results could help. (I think someone else mentioned that on another ticket.)
The original mockups I did were based on the A/B test UI. They weren't intended as UI examples, but rather as possible kinds of information to show. People weren't understanding the descriptions as well as I'd hoped, so I added pictures. :)
The current cross-wiki results are pretty much the same, too.