Home » forums » LtU Forum

The Worst Monad Tutorial... Except For All Those Others.

Excuse the tongue-in-cheek title of my recent post detailing my adventures in writing monads in C# 2.0. I believe I've finally nailed down the problems I was having, so thought I'd share. I'd welcome any corrections. :-)

By naasking at 2008-01-19 02:38 | LtU Forum | previous forum topic | next forum topic | other blogs | 1831 reads

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

Type-safe general monad interface in C#

Well, almost type-safe. It uses casts, but only for the monad implementors accessing internal state. The interface is very simple and strict, so it should be hard to make a mistake. Details here.

By naasking at Sat, 2008-01-19 20:36 | login or register to post comments

Interestingly, it is

Unfortunately not, because C# cannot abstract over type constructors.

Interestingly, it is possible to simulate a very restricted class of higher-kindedly polymorphic functions in C# by adding sufficient extra type variables and related constraints. For example, one could encode 

class Container c where
  fst :: c a -> a

fstfst :: (Container c, Container d) => d (c a) -> a

by

interface Container<A> { A Fst(); }

A FstFst<DCA,CA,A>(DCA x) 
  where DCA : Container<CA> 
  where  CA : Container<A>

But clearly this pattern only applies to type signatures where the higher-kinded type variables are not applied to different arguments within the same signature. One could not encode something like, for example:

  join :: (Monad m) => m (m a) -> m a

so the class of functions that can be encoded is very limited, and as usual with C# there's a fair bit of syntactic noise.

By Steven Ramsay at Wed, 2008-01-23 10:27 | login or register to post comments

Really higher-kinded?

It seems to me that what you've simulated is much closer to this:

class Container c a | c -> a  where
  fst :: c -> a

fstfst :: (Container c a, Container d c) => d -> a

Am I missing out on a subtlety?

By Daniel Burrows at Wed, 2008-01-23 15:55 | login or register to post comments

No, I don't think there is

No, I don't think there is any subtlety. The restricted class of functions are precisely those which can be written in the way you demonstrate using fundeps in (non-standard) Haskell.

By Steven Ramsay at Wed, 2008-01-23 17:23 | login or register to post comments

I believe you can encode

I believe you can encode join using the technique in my follow-up post.

By naasking at Wed, 2008-01-23 16:01 | login or register to post comments