Patterns of generic programming

Jeremy Gibbons describes seven patterns of generic programming, where each pattern is viewed as a different kind of parameterization. For more information, refer to Datatype-Generic Programming (available at http://www.cs.ox.ac.uk/jeremy.gibbons/publications/dgp.pdf).

Patterns 1 and 2 – functions

Functions parameterized by values are more general than functions with hardcoded values. This is the most simple kind of generic programming.

Functions parameterized by functions (higher order functions) represent a more powerful form of genericity.

Pattern 3 – polymorphic types and functions

Polymorphism can also be viewed as a pattern of generic programming. For example, when we parameterize types by other types (type polymorphism):

  Tree a = Leaf a | Node a (Tree a)
  – is more generic than
  TreeI  = Leaf Int | Node Int TreeI

Or when we parameterize functions by types (function polymorphism):

  f :: Tree a → a
  – is more generic than
  g :: String → String

Pattern 4...