Thank you, Robin, for starting this ball rolling. I have just spent some time reviewing the existing use-cases, and "liking" those that seemed important for us. It soon became clear, however, that many of the use-cases there are germane to the API, others to the uderlying storage layer, and others are specific to applications that are built on top of Fedora and which may themselves have changed a lot in recent years.
Given that we now have the API as a starting point, would it make sense for us to start a new set of use-cases specific to the task of implementing the API over a different storage layer? I wouldn't want to re-invent the wheel unneccessarily, but it feels like that might make it easier focus on the things that are up for discussion with the next major release.
Best,
Josh