On definitions of bounded variation for functions of two variables