Abstract We reexamine the approximate formula for the path difference in the double-slit interference experiment, establish its limits of applicability from a geometric perspective and discuss its validity for the two-source interference of acoustic and surface waves.