The "spiral" model was documented in 1986 and in common practice long before Extreme Programming hit the shelves. I never saw a waterfall-like process in use in the 90s except in the DoD. It was never more than a strawman for XP to beat up on. (Edit: On a hunch I just looked up when scrum originated: 1986 as well.)
Some other arguments that show that XP wasn't as new as the book argues to be:
- I doubt Lisp, Smalltalk, Forth, time sharing Basic and similar tools were invented for the waterfall model, and all were available in the '70s.
- Weinberger advocated egoless programming in 1971.
Also, looking back at the 1960-1975 timeframe, nobody would have accepted XP as a methodology. I can see the discussion "what do you mean by 'if it turns out we need more memory we simply move to a larger computer'? Ordering one will take months, it will cost us thousands of dollars a month, and we will have to build a new computer room. And no, it is not acceptable if that accounting program you are writing turns out to be a glorified checkbook because that is what you can cram into the machine."
Also, it wasn't as easy to get a MVP because there were so few libraries to build on. IBM would happily license you their 'database' software or their time sharing software, but you better be sure that you needed it, because you paid for it by the month. But it just wasn't possible to work on the idea "don't worry about the 3D graphics. We'll eventually pick a library, but that can wait."
A devil's advocate would argue that the XP book just was a good summary article with great marketing that due to pure luck appeared at just the right time.
