With regard to quota management: the problem with trying enforce a "hierarchy" quota as opposed to a "single folder" quota is that it may be the case that not all the subfolders are replicated locally. The server can't compute the size of the content in a folder which isn't replicated locally. (This is the same excuse for why we don't support deep searches - the results would be incredibly confusing.) We don't want to store the computed size of the content as a replicated property of the folder itself as that would create a rash of unnecessary hierarchy replication every time any content change happened.
To implement this feature, we'd require full-time connectivity among server so they could each query each other about pertinent information. The PF system doesn't otherwise have this requirement, and implementing it would be a substantial change in architecture.