A few remarks on the supremum of stable processes