[Chattered]

> I'm still looking for someone who can present a cogent explanation of monads using only Python.

Personally, I don't do monads in any other language than Haskell. I find monads painful even in Ocaml. I'm pretty stupid as a programmer. And --- especially when doing things such as monad transformers --- I need lots of type checking and lots of type inference to help me over my mistakes (which are pervasive).

The cool thing is that when I've written these things called "monad transformer stacks" (see Real World Haskell for what I guess is the best explanation for those), I find that when my code gets through the type checker, it really does just work. It works, even when I've lost any sense of what my code is actually doing underneath. Without the type system, I couldn't have any confidence of this, and the runtime bugs you can potentially get with monads are so subtle that I would be terrified to use them in a language like Python (note again, I'm stupid, and YMMV).

The other reason why I wouldn't do monads without types is because, to me, monads are the type. The semantics of a monad is pretty much just that type (plus some crucial axioms). Types in Haskell are much more expressive than in most languages. The way "operator overloading" works in Haskell means that by specifying types, you are often generating runtime code (think C++ templates, but not for too long). That's why you see more type annotations in Haskell compared to (say) Ocaml. Though if you want to do the same thing in Ocaml, you'll be knee deep in the even more verbose world of these things called functors.