I don't see your point. In every case where you depend on filenames to identify blobs of data, case-sensitivity (actually byte-sensitivity) is the only behaviou

I don't see your point. In every case where you depend on filenames to identify blobs of data, case-sensitivity (actually byte-sensitivity) is the only behaviour that a filesystem can reliably implement. Once you allow cases where two different byte sequences must be considered equal depending on configuration, your filesystem can no longer work reliably. Case sensitivity may bother users sometimes, but the filesystem is not the right place to solve that problem.