That's a nice, succinct way to put it. Software that doesn't interact with the world can be finished. Software that does ends when the world does.
I suppose then that a benefit of pure functional programming is that you can statically analyze whether a given piece of software is plausibly "done" or simply unmaintained, depending on whether and to what extent it uses monads.
I suppose then that a benefit of pure functional programming is that you can statically analyze whether a given piece of software is plausibly "done" or simply unmaintained, depending on whether and to what extent it uses monads.