Does this kind of advice apply to all research or specifically mathematics?