Postscript
==========


Summay: What we have done

  We've been able to sample the capabilities and programming style used in
  functional programming.

     strict, impure in SML
     lazy, pure in Haskell

  We covered virtually every aspect of SML, but a skilled SML programmer needs
  to master technique and the libraries and tools, and gain experience 
  programming at scale.

  
  With Haskell, we have covered some of the main ideas:

    - type classes  (basic and constructor classes)
    - monads
    - lazy evaluation

  But there is lots more:

    - first-class polymorphism
    - programming with multiple monads
    - generalized abstract datatypes
    ...

  making Haskell a much more complex language than SML.

  
  Both SML and Haskell are statically typed, with type systems based on
  the Hindley-Milner type inference algorithm, yielding polymorphic types.
  We explored a realistic implementation of Hindley Milner (typecheck2).

  To support "interpreted types", SML uses its module system, while
  Haskell uses type classes.


Unfinished business.

  Exploring the deeper levels of Haskell (fancier types)

  Further exploration of concurrent functional programming

  Application areas like functional reactive programming

  Connection with logic and proof (proof assistants like Coq)

  Specification, testing, and verification.

  Dependent types (e.g. the Agda language)

  Contrasts with OOP (the dominant "paradigm")