I read a Kindle-only book several years ago, Managing the Mechanism, which made an extremely simple point that lays these simple-minded methodologies quite firmly in the dust:
Computer software is a mechanism. It is an automaton, a software machine which will do everything that it does unattended, automatically. Yet it is a mechanism of fearsome complexity, composed of literally millions of parts, any of which can and do interact with each other. The team that built it will put it on the field and push the Start button, leaving it to play the entire game ... alone. There is nothing like unto it in all human experience.
The book goes on to argue that “project-management principles” simply fail to apply to software projects, as they might for any other type of project, for precisely this reason.
