Freeman Dyson said no scientific (or, I think, technological) project should last more than about five years, because the context under which it was conceived changes too much in that time.