What is the "monad-style behavior" they talk about relating to enums here? I really don't understand the connection, and nothing in the rest of the proposal makes any mention of it.
The draft there has a feature that the posted RFC does not. Associated types. Without the associated types you can't have the monadic style behavior which is why I was similarly confused by that wording there. Associated types elevates this from "Oh look a C enum" to actually useful ADT's like in Rust or the ML family of languages. I wonder if the draft you linked will make it into the official rfc or not.