ok. I've got surprised then (Workflow transition on defaultPage is propagated to its container · Issue #3932 · plone/Products.CMFPlone · GitHub)
I would point out that Plone is great in many ways but there are a subset of "features" that row against it's greatness.
There are "complex use cases" doable only by a administrator/developer (and this is expected in all the framework) and other "complex use cases" doable by a simple editor which is GREAT (create custom type, define a whole workflow, content rules, collections, tiles/block, url aliasing, caching rules, translations...)
There are "simple use cases" doable by anyone (and this is expected in any framework) but, when there are "simple use cases" that needs to be done by an administrator/developer we have a problem.
Plone has problems in how it manages, let's say, "top folders" in medium to big portals.
Renaming an item it's a simple thing, a simple/trivial use case.
Publishing/unpublishing an item, the same
Try to rename or publish/unpublish a top folder in a big portal... there are high probabilities that it would not converge due to conflict errors, and the task needs to be delegated to an administrator/poweruser/developer.
In this particular scenario I've got surprised because I unpublished a simple default page, and Plone took over 30 seconds to commit the transaction. Then I realized that it's parent folder got revoked too and Plone, in these cases, fire a recursive security reindex. That folder contained thousands elements.
Renaming a top folder (it's another scenario) it's even worse because Plone does a full reindex (of all indexes) of the whole subtree. not a joke.
This is not meant to be a rant as it might seem but this is just food for thought.
alessandro.