Does IETF have any way of enforcing such policy?
(Yes. Peer pressure. But an RFC which turns out to be infeasible to implement or wildly speculative or what-have-you doesn't matter, because there won't be any peers pointing out that you ought to do it this way...)