List notation

We are used to the list notation [0,1,2,3]. However it is syntactic sugar for (0:1:2:3:[]). By using the syntactic sugar, we often miss the benefits of the direct notation.

• A trailing colon is like a terminator.

0 :
1 :
2 :
3 :
[]

Thus it is more theoretically sound and easier to edit.

• You can easily mix elements and lists into a list by appending the corresponding operator in each line:

[1,2,3] ++
4 :
listA ++
5 :
listB ++
[]

• You can insert elements or sub-lists conditionally.

infixr 5 ?:, ?++

(?:) :: (Bool, a) -> [a] -> [a]
(?:) (b, x) = if b then (x:) else id

(?++) :: (Bool, [a]) -> [a] -> [a]
(?++) (b, x) = if b then (x++) else id

list =
   [2,3] ++
   (x==5, 5) ?:
   (x==7, listA) ?++
   []

• You can construct a singleton list with a section of the colon operator:

  (:[]) :: a -> [a]
  

  .
• You can prepend an element to a list:

  (x:) :: [a] -> [a]
  

  . E.g.

  iterate (' ':) []
  

  creates a list of blank strings with increasing size very efficiently.
• You can adapt this style to other list-like data structures, e.g. a list of elements with alternating element types. See e.g. event-list.

data Alternating a b = Alternating a [(b,a)]

infixr 5 /., ./

(/.) :: a -> [(b,a)] -> Alternating a b
(/.) = Alternating

(./) :: b -> Alternating a b -> [(b,a)]
b ./ Alternating a bas = (b,a) : bas

example :: Alternating Bool Int
example = True /. 0 ./ False /. 1 ./ True /. []

See also

• Haskell-Cafe thread: Comma in the front