Re: [Haskell-cafe] Alternative IO

On Fri, Jul 10, 2009 at 10:35 AM, Wolfgang Jeltsch wrote:

... Hello Cristiano,

I fear that this instance doesn’t satisfy required laws. As far as I know, the following equalities should hold:

   (*>) = (>>)

   f *> empty = empty

   empty <|> g = g

This implies the following:

   (f >> empty) <|> g = g

But this wouldn’t hold with your instance. (f >> empty) <|> g would cause the side effects of f and of g, while g would (obviously) only cause the side effects of g.

If empty would be a real empty, it would have to undo the effects of previous actions (like f above). So an Applicative instance makes sense for STM but not for IO.

Thanks. That's an argument :) -- Cristiano