Many of the trad engineers I interviewed said that waterfall wasn't considered best practice in their trad field, either. Everybody does iterative development with many concurrent feedback loops

Also the people who originally advocated Waterfall (Boehm and Royce) quickly reversed course and started recommending against it within maybe a decade, but by that point everyone had already fixated on the original unrevised idea, or at least what the original idea sounds like

And I suspect you know this, but Royce’s paper actually argues against Waterfall (it says that even a modified Waterfall with some feedback loops is “doomed”).

@Morendil has a really great investigation into that paper, and makes a good case that it was, in fact, pro "Waterfall":

Yes, and I think he’s mostly right; it certainly doesn’t support iteration. (And what it *does* propose is not at all sensible.) But when people say “waterfall” they’re usually referring to a strict phased process a la Royce’s figure 2, which he clearly thinks is rubbish.

All of which is to say: Royce doesn’t quite deserve the scorn that was heaped on him for many years, nor the accolades from the revisionists who want to paint him as a proto-agilist.

...you know, I have **no idea** what people thought about Waterfall in the 80's and 90's. I only know about how Agile _compares itself_ to Waterfall. I'm suddenly VERY curious about the history here.

I do know that Booch, Rumbaugh, and Iverson all considered their object-oriented methods as processes distinct from "Waterfall", which makes it ironic that their magnum opus is so heavily associated with what we now call "Waterfall"