Overview

Monad Tell

MonadTell is used to return additional non-output data that is generated during a computation. It's default implementation is WriterT.

Since we can only return one object and we want to return something in addition to the output, we'll need to return a Tuple that wraps the output and additional data. In cases where we already have non-output data and need to "store" another alue of non-output data, we'll need to combine the two together, which implies a Semigroup. Lastly, to implement Applicative, we will need an "empty" value of that data, which implies Monoid.

Putting this into code, we get this:

             -- w               m     a
newtype WriterT non_output_data monad output =
  WriterT (monad (Tuple output non_output_data))

-- Pseudo-syntax: combines the type class and instance into one block
class (Monoid w, Monad m) <= MonadTell w (WriterT w m) where
  tell :: w -> m Unit
  tell w = WriterT (pure (Tuple unit w))

Do Notation

Since tell returns an m Unit, which will be discarded in do notation, we'll only be writing:

useReader :: Reader NonOuputData Output
useReader = do                                                          {-
  unit <- tell nonOuputData                                             -}
          tell nonOuputData
  -- without indentation
  tell nonOuputData

Monad Writer

MonadWriter extends MonadTell by enabling a computation's non-output data to be

  • appended via tell and then exposed in the do notation for later usage (listen)
  • appended via tell after it is modified by a function (pass)

Derived Functions

MonadTell does not have any derived functions.

MonadWriter has two:

  • listens: same as listen but modifies the non-output data before exposing it to the do notation
  • censor: modifies the non-output data returned by a computation before appending it via tell.

Laws, Instances, and Miscellaneous Functions

For their laws, see

For WriterT's instances:

To handle/modify the output of a writer computation: