I figured this 1994 paper would be relevant to the discussion about finished software, "Software Aging" by David Parnas. Read it for a software engineering course where the prof railed against the assumptions underlying the notion of finished software.