Tony Morris

Is this supposed to illustrate a contradiction? I don't see any.

Or perhaps I didn't explain myself correctly. `x = readLn; [x, x]`, by itself, will NOT perform any side effects. `[x, x]` is an expression of type `[IO String, IO String]`, and that's all it is.

By itself readLn in Java will not do anything either. Because a program that isn’t run doesn’t do anything. As soon as you *run* it, no difference. Even the expectations and reasoning is exactly the same as in any other labguage with HoFs

I've gone through this before but ok: Java's `readLn` is equivalent to Haskell's `readLn`. And Java's `readLn()` is equivalent to Haskell's `unsafePerformIO readLn`. Difference is, you write readLn() all the time in Java, but you never write `unsafePerformIO readLn` in Haskell.

You *can* write pure FP in Java, but you'll be fighting the language every step of the way.

Why? Ok, not Java but JS. Or Erlang. Anywhere you have HoFs basically (so, even Java)

I can't speak for erlang, but java lacks even the most basic foundations for pure FP. You can't even combine a nullary function (Supplier<A>) with a unary function. I suspect js is equally handicapped.

A lot of really smart people have put a lot of effort into building those foundations, and IIRC the conclusion was that it was like putting lipstick on a pig. Maybe @dibblego can fill you in on that.