In 1970, a software engineer named Winston Royce published a paper that would become one of the most influential and most misread documents in the history of software development.

The paper described a process in which software moved through sequential phases — requirements, design, implementation, testing, deployment — each completed before the next began. The diagram showed these phases flowing downward, like water over a series of steps. It was called, eventually, waterfall. Royce did not call it that. More importantly, Royce did not endorse it.

The paper’s actual argument was almost precisely the opposite.