is self-answering questions on PhysicsOverflow generally encouraged?
Probably nobody would object to the scenario, where somebody asks a question of which he does not know the answer when posting it, subsequently finds the solution/explanation himself, and then posts this as an answer too.
Another potential use of self-answering, suggested by Arnold Neumaier, would be to give a clearer expositions of material that is presented in the literature only in a more obscure way, which would be rather an answer and the question would have to be "constructed".
What do people here think about this?