Your second-to-last sentence is part of what I am trying to say. Ikegami (as I understand his post) proposes that the ordering of keys intrinsic to an array isn't a difference to a hash because a hash can simulate that (and because internally there is an ordering in a hash anyway). My argument is that
b) if we argue that a hash is as ordered as an array because we can make it so, then all arguments about differences of data types become meaningless because every data type can simulate every other data type.