That seems unfair.  Please reconsider. (your opinion carries a lot of weight and can be self-actualizing!)

There is certainly a wariness of fragmenting experiments that don't have the promise of bringing new ideas back into our core tools and sites. It's great to borrow from & integrate w other projects, and the integration needs to happen. We need enough gravitational attraction in the core to tie things together.  

If we had a strong persistent vision for how to support multilingual discourse on our projects, and someone leading its design was warmly engaged here, pointing out how this contributes to the ongoing work, most of the stated concerns would go away.

 

  One-click translation is important.  Talk page sections (and flow) should have it.  There are solvable details.
  Automatic link-expansion can be handy.  Talk pages should have it as an option.  There are solvable details.

  The ability to sort a list by voting is important. (People do this laboriously on wikis all the time!) We should have an in-band solution. TASD.

Let's practice fast clear templates for experiments:

 1. Social templates for saying 'we want this', proposing simplest workable experiments 

 2. Clarity of maintainers and decision making. Who (in each area) can say 'we plan to have this', update roadmaps or plans of record, allocate time, iterate onp experiments?

 3. Technical templates for trying new tools in a way that informs and improves our core, and is persistently reusable. (We have tried many discourse instances. What is the recommended way to synchronize it with a wiki?) 

 4. Giving shout-outs to existing work. Underrated + uplifting. We have community scripts or feature requests that have tried to add most of these features to MediaWiki itself. Naming them and their outcomes can highlight what we still have to learn from new implementations.

SJ

(PS. I like this experiment & it motivates MW improvements. But for those who might not immediately think of what we /lose/ by moving a q&a to a forum, off the top of my head:

 ~ unified recent changes

 ~ unified notifications

 ~ nestable threading

 ~ coherent archiving

 ~ easy wiki-linking and backlinks

 ~ easy transclusion onto other pages

 ~ the abstraction of talk pages (or other annotation)

 ~ translation flow for non-automated translation

)