Functional Programming in C#: Classic Programming Techniques for Modern Projects

WHAT’S IN A MONAD?

In Haskell, Monad is a typeclass. Here’s what it looks like:

class Monad m where

  (>>=) :: m a -> (a -> m b) -> m b

  return :: a -> m a

Granted, at a glance this doesn’t explain much. But the important thing is to recognize that a monad is a particular typeclass, in other words, Monad describes a particular class of types. It is the class of types that support the two rather cryptic functions (>>=) (this is usually pronounced “bind”) and return.

To illustrate the requirements that are declared by the Monad typeclass, here’s an implementation of a C# interface with a very similar set of two functions:

public interface IMonad<A> {

  IMonad<A> Return(A val);

  IMonad<b> Bind<b>(Func<A, IMonad<b>> g);

}

That looks much simpler! The Return function takes a value of type A and returns an IMonad<A>. It wraps the value in the monad, so to speak. The Bind function takes a delegate as an argument, which can receive a value of type A and returns an IMonad<B>. Bind itself also returns an IMonad<B>. In reality this interface is usually not used because the formalization it provides is regarded as too strict. For example, the process of wrapping a value in a monad, as described earlier by the Return function, might be implemented as a constructor call in object oriented C#. The two functions also could be implemented as static methods, perhaps extension methods, elsewhere, instead of being instance methods in the type itself.

To summarize the theoretical part: a monad ...