I first saw the "Release Early, Release Often" in Eric Raymond's great book "Cathedral and the Bazaar". From what I understood at the time, he was pointing out the the actual revolution of Linus was not in fact the kernel, but the methodology. The "promiscuous" programming practice of Linus was based on the Release Early, Release Often principle, and Raymond later proves that theory with his Fecthmail controlled experiment.
I personally think it is no coincidence that the "Rational Process" recommends the Iterative and Incremental method by the late nineties as if it were a natural evolution of _their_ programming practice. IMHO, they were just trying to formalize what Linus had invented almost a decade earlier, but since "Open Source" (by OSI definition) was not coined yet, and it's practices were not accepted in the "Enterprise"[ 1 ], Rational took the opportunity and re-discovered warm water[ 2 ].
Notes:
[ 1 ] This is a term I truly despise, specially in the words of Gartner and alike, and in things like J2EE which a far from being anything "Enterprise".
[ 2 ] Sorry if you might not cache the joke here, but it's a Venezuelan expression that refers to a foolish person that thinks he has discovered something new, when it's actually something well known.
| [reply] |