The rest of the REPL is just about as passive-aggressive, but not quite as loud.
no, I don't know why you can dynamically assert facts but not rules.
It kind of defeats the whole point of having a REPL at all.
This kind of silliness is why Prolog doesn't quite cut it for interactive exploratory programming.
Maybe the point is that the negation by failure part fails if you run it in *parallel*, which is what most of 80s Prolog / database research was about.