Topic on Talk:Reading/Web/Projects/Related pages/Flow

A few weeks ago I already tried to tell you that: Please, please stop doing the same mistakes again and again.

1. This feature is pure clickbait. It serves zero purpose other than tricking people into not leaving the page. But why? Really, I'm an active Wikipedian for way more than a decade, but I can not think of a reason why we, as a movement, would want that. Since when is the goal of a free encyclopedia (!) to make people stick to a site as long as possible? Who came up with this goal, when and why?
2. Where is the community poll that asked for such a feature?
3. Why is the feature in the article? Why does it look like it's content? Why are you messing with the content again? I strongly believe this is one of the top reasons why so many WMF projects caused such an insane pain (most notably Media Viewer and Typography refresh): Because you, the WMF, are not supposed to, not even allowed to touch the content of Wikipedia articles. There is no way the feature can stay where it is. Go and move it out of the content. Make it opt-in. Make it possible to kill it per user, per user group and per wiki. Start polls and kill the entire feature when more than 100 of the most active Wikipedians dislike it. Burn it with fire and move on to features people actually want.
4. The links are barely more than random noise, as I and many others already tried to explain in much more detail. And no, there is no way you can solve this algorithmically. Creating useful content that serves a purpose is a job that must be done by a human. The only thing you could do is to turn this into a tool that suggests links that are currently missing in an article.
5. How to disable the random noise at the end of each article in the mobile app(s) when I, as a reader, find it distracting?

This is the first time I feel the need for an ad blocker on Wikipedia. This scares me like hell.

@TMg "Make it opt-in. Make it possible to kill it per user, per user group and per wiki" - if I see it correctly, it's currently just an opt-in beta feature anyway, right? I assume that it will - after past experiences with various features - not be introduced without a clear community consensus in the respective project.

I assume it will – after past experiences – be enforced on all smaller wikis that do not scream loud enough.

Things are a lot better now. The big events were quite some time ago, and the people who were in charge are gone. In fact you'll like this quote:

"The WMF... does not want to push it on Wikis that do not want it."

Mostly you're preaching to the choir here. This feature produces irrelevant results, wasn't asked for, et cetera.

The feature is the way it is because it's imported from the mobile site, which has made a habit of reformatting Wikipedia without talking to the community about it, in part because they're trying to compete with third-party Wikipedia apps that also pander to readers without any regard for the editing community. Did you know that the mobile app doesn't even support Special:Watchlist?

But I digress, because I'm annoyed about it too. Expressing our frustration isn't particularly helpful, because at heart everyone's on the same side here: we all want a better experience for both Wikipedia readers and editors, and it's really important that we continually assume good faith here.

Some of the more productive discussions here have focused on ways to take the best elements of this feature and integrate it into functionality that can be more useful. For example, perhaps "See also" sections would benefit from PageImages-based thumbnails. How else could this feature be repurposed into something of actual value? It serves our opposition to the present feature better to help developers pivot it into something constructive.

I did that in my initial post and repeated it in my post above: Make it an editing tool that suggests missing links and categories. I find this obvious, based on the simple observation that every bigger, stable community constantly refuses bot created articles.

Discussing other usages of PageImages is not helping here, I believe. This is an other extension. "Related pages" uses thumbnails provided by PageImages, but nothing I criticise would change without them. Or the other way around: When you kill the algorithm, there is nothing left to discuss here.

Or even simpler: "Related pages" aims to solve an already solved problem. We call it "link". We link to other articles when it makes sense and do not when it doesn't.