My initial view of the constraints was that they were extra requirements which your implementation should accommodate for, other views I have heard are they are for information purposes only so that your implementation does not handle unnecessary test cases.
I would have thought implementing these in the code handles the latter in the above. Any thoughts?