One upside is that it hopefully prevented developer to ship half-baked software that rely on filename and can't handle duplicate name gracefully.

You can't prevent collisions (multiples sources/counter reset/date reset, etc). So it's actually nice to have an unforgiving standard that will bite you if you make unfounded assumptions.