> What makes you think that? Haskell is the only language outside of the proof assistant family to avoid the temptation to conflate evaluation with execution.

> What makes you think that? Haskell is the only language outside of the proof assistant family to avoid the temptation to conflate evaluation with execution. It’s pretty clear in retrospect from looking at old papers on e.g. lazy list IO and old mailing list threads that modern monadic effects were born out of “well shit, Haskell makes us actually deal with this problem instead of doing what every ML language does with ()->” > Please elaborate. An object that is not referentially transparent is not a function. “Functions” in most languages are not necessarily referentially transparent, and even if they happen to be in a given case, it’s usually clear to neither the programmer nor the compiler.