It's not that far off, if you think of side-effects / any magical processing happening at evaluation per bind-function (semi-colon ;). It's so because monads f

It's not that far off, if you think of side-effects / any magical processing happening at evaluation per bind-function (semi-colon ;). It's so because monads follow the exact same pattern by definition, but the different types of monads just inject some additional behaviour and constraints as well. Different patterns from this are no longer monadic. The "magic" part of this is why different types of monads don't compose well. Of course, this is overly simplified but might help for understanding.