I don't think we should rush the naming on the eve of a feature freeze.
As I mentioned above, this mainly affects the new API, with a very limited impact on the old API, so I think we should be fine. I'd rather make a couple renames in the next release than to withhold the dozens of renames in this patch that really make the new API much cleaner.
how the patch retains both the old and new hierarchy
The same instances of SolrInfoMBean-s are registered under the old names and new hierarchical names. This way we preserve the back-compat, and at the same time we provide an alternative, improved view of the same information.
though perhaps at the base level a generic simply HANDLER (I dunno).
Yeah, naming is hard On one hand it needs to be specific enough to be informative, on the other hand it needs to be generic enough to limit the total number of categories, and on the third hand if it's too generic then it becomes meaningless.
I discussed this with Shalin, we considered and rejected HANDLER, PLUGIN and the like as too generic, after all all these components are handlers and plugins. In the latest patch I moved quite a few components from QUERY to other categories, leaving only those that actually support query-response interaction with users.