The JIRA has not gone very far because it was hard to do (I just updated my own findings there). solr-start, based on similar principles is up and running, serving user needs and getting good views. And it is finding components that were missed both by Reference Guide AND Release Notes, exactly because it relies on shipped code alone. And, things like cross-reference to usage example, is not something that Reference Guide can maintain by hand with reasonable effort.
The V2 API might be the magical unicorn when it actually lands, but - as my own comments there indicate - I could not make heads or tails out of it in a meanwhile. So, let's focus on reality, especially the reality for 6.x.
As open-source Solr developers, we are combining 3 roles that are often separated in commercial organizations: Developers of new - blue-sky - features (master), maintainers and developers of the current features (6.x currently), and tech-support for whatever was already released (6.x.y, 5.x.y). Your suggested approach is about blue-sky which is great, but not sufficient. Let's just take care of these other aspects too, so other people don't have to.