Hi Christian,
I do not know any reason why BOPF should check for EDIT authority on a subnode when deleting the parent instance. I made a quick check with a test BO in BOBT: I debugged the deletion of a ROOT instance with a single ITEM and I did not observe an authority check for activity EDIT. I just saw two authority checks for DELETE ('06') for the ROOT and ITEM node, respectively.
However, I did not test an extended BO.
Please analyze the call stack leading to the unexpected authority check. Is there any BO-specific logic involved which does not belong to the BOPF framework? Otherwise feel free to open an incident, so we can analyze your situation.
Kind Regards,
Ivo