There’s a big difference between testing whether someone already knows something vs. whether they can figure it out on the fly.
The tortise-and-hare algorithm conflates the two: if you’ve already heard of it, it’s quite simple to implement. If not, it takes a flash of insight to consider having two pointers chase each other through the list, one moving multiplicatively faster than the other. I’d bet that flash didn’t come to Floyd in the first 15 minutes he spent on the problem—-which certainly wasn’t under pressure in an interview.
Sure. The problem with the argument I wanted to highlight wasn't the puzzle nature of the solution, but the fact that the OP chose to instead focus on the research angle. I mean CSP, Actor Model, Lisp have strong foundations that took a lot of research... as did everything else from Type Systems to Linkers to Compilers. That said, I get your point too, I might have misunderstood OP.
The tortise-and-hare algorithm conflates the two: if you’ve already heard of it, it’s quite simple to implement. If not, it takes a flash of insight to consider having two pointers chase each other through the list, one moving multiplicatively faster than the other. I’d bet that flash didn’t come to Floyd in the first 15 minutes he spent on the problem—-which certainly wasn’t under pressure in an interview.