Just preparing for new optimization to be used in all traversal implementation. Should be no measurable difference.