So, right now every time I start contributing a campaign, it starts at the first file -- frequently one that already has captions and depicts statements -- which may be useful when you want multiple sets of eyes on a file, but isn't so great when you are trying to get to "completeness"-- could you add a random start for the set? I know this worked well on https://tools.wmflabs.org/citationhunt/
Hi @Astinson, thanks for the feedback!
Our plan was to just randomise the order of the images initially, which I guess gives a very similar effect to random starting point :)
Actually just completed so should be live in a day or two - T224019
I feel like we really would benefit from adding more advanced logic though. Starting with images which have been viewed the least, or have the least statements?
served based on a "view score" that would take into account
- not viewed images (or rarely viewed) get higher score
- completeness (image with no depicts and no captions gets a higher score than those with several depicts)
- viewed by many users but no improvement provided would get lower score