What sort of surprised me as I shopped my copy around and showed people the python-add-one-factor example is that even programmers I consider experts didn't realize how quickly the conditional cascade blows up. It's easy to miss. And I looked around for people to mention this in blog posts, but almost no one does. So I feel like there's a bit of life left in the old fizzbuzz yet.
I think the non-blown cascade is exactly what makes fizzbuzz aggravating. With three noises, the linear solution clearly dominates. With two, the exponential is actually shorter -- but feels unclean.
I've been programming some game servers, and I have the same problem with guaranteed two-player games; I feel dirty hard-coding the logic to assume two players, yet making it general enough for N players makes it absurdly more complex for no gain, which is a net loss.
(And yes, before anyone pops in, these are guaranteed two-player games. Of all the rules of the games in question which have changed over time, that is the one rock-solid constant which will not change in this application.)
What sort of surprised me as I shopped my copy around and showed people the python-add-one-factor example is that even programmers I consider experts didn't realize how quickly the conditional cascade blows up. It's easy to miss. And I looked around for people to mention this in blog posts, but almost no one does. So I feel like there's a bit of life left in the old fizzbuzz yet.