In my observation, roughly one out of every five software engineering fads/trends "sticks", meaning that 4 out of 5 times the fad/trend fades into oblivion or into narrow niches. Do others agree on this? Sure, there are a lot of grey areas; this is just a general rule of thumb, kind of the MooresLaw version of tech fads.
Note that "sticking" does not necessarily imply that it takes over, but may merely become one tool or technique among many. The "takes over" rate is more like one in thirty, by my thumb.
--top
kind of the MooresLaw version of tech fads.
Perhaps it should read "kind of an EightyTwentyRule of tech fads."?
And the distinction between "sticking" and "takes over" seems like another round of the EightyTwentyRule.