[alganet]

He's right.

When the author says over-engineered stuff is future-proff and all of that, he's implying _successful_ over-engineered stuff. Same for under-engineering.

Of course most over-engineered projects fail. Most under-engineered too. That's why we learned all those things people are parroting in this thread (we parrot because it's useful).

However, _some_ over-engineered projects will succeed. Those will most likely present a nice combination of future-proofing, scalability and reusability at the core of their success. Some under-engineer projects will iteratively limp their way into excellence as well.

He's not asking you to think of scalability and such, on the contrary, his recipe for following the middle way is quite reasonable (although a bit generic, like these kind of things always are).