If we want a useful, testable theory that "complexity" causes "software collapse", we need to define both independently from each other and find a way to meassu

If we want a useful, testable theory that "complexity" causes "software collapse", we need to define both independently from each other and find a way to meassure both. Not that I doubt that there is something to be learned here (see for instance the famous Mythical Man-Month). But without a better theory, we are still poking around in the dark. (C64 pun intended.)