Oh, well I can agree with that.

You can like the idea behind functional programming while believing that any application is in the end about side effects and therefore a purely functional application impossible.

It's a bit of a tangent, but if you're doing something completely deterministic and non-interactive, like computing a digit of pi, it's great in practice as well. I use Haskell semi-regularly for that kind of thing.

You could argue printing the output is a side effect, but is a side effect followed by termination really "side"?