@TBolliger: It sounds like the blocking task (T59617) isn't going to happen any time soon, so I recommend declining this task. We can use some other proxy of article prominence like size of the article or number of incoming links instead.

We can use some other proxy of article prominence like size of the article or number of incoming links instead.

Or pageviews!

I'm not sure why T59617 is a blocker. What is stopping us from using the API for getting that information? I can't imagine that it's a lot more work.

@Niharika: We could use the API, but what do we do when an image in used in 100,000 pages? Do we want to make 100 API requests and compile all the data together? You're right that it's not that complicated, but it also feels like overkill for such a low priority edge case. Even if we used absolutely no logic at all and just posted notices on the first 10 pages, that would probably be fine. Images used on lots of pages are rarely deleted and even posting on 10 random pages is likely to get enough attention to make sure it's discussed. Relying on the API is also probably more fragile than making a query (as backwards as that may be). That said, I don't really have a strong opinion on this.

@kaldari How common do you think it will be for an image being used on 100k pages to be deleted? I don't think that will be a common scenario.

@Niharika: Almost never, but I imagine we would have to code it to accomodate that case in case it ever occurred (if we were assessing number of watchers). Regardless, I don't care too much one way or another. I'm fine with whatever y'all want to do on this.