Personal tools

Terminator vs. separator

From HaskellWiki

Jump to: navigation, search

There are several concepts for notation of sequences of elements. Usually, we don't think much about it. Programming languages provide different schemes, often different schemes in the same language, we are used to them, and no scheme seems to be better than the other one. However, there are differences and good reasons for preferences.

1 Terms

  • Separator: There is a symbol between each element. This is what the functions
    Data.List.intersperse
    and
    Data.List.unwords
    generate. In Haskell language, the following syntaxes allow separators only:
    • list sugar:
      [0,1,2,3]
    • tuples:
      (0,1,2,3)
    • type class constraints:
      f :: (Show a, Ord a) => a -> a
    • declaration of named record fields:
      data T = Cons {a :: Int, b :: String}
    • declaration of data constructors:
      data T = A | B | C
  • Terminator: There is one symbol after each element.
    • list notation using infixes can be considered an example:
      0:1:2:3:[]
  • Liberal choice between separators and terminators:
    • export lists:
      module A(a,b,c) where
      and
      module A(a,b,c,) where
      (and
      module A(a,b,c,,,) where
      ...)
    • import lists:
      import A(a,b,c)
      and
      import A(a,b,c,)
    • let syntax:
      let a = 'a'; b = 'b' in ...
      and
      let a = 'a'; b = 'b'; in ...
    • do syntax:
      do a ; b
      and
      do a; b;
  • Initiator? Theoretically it would also be possible to introduce each list element with some symbol. However, I don't know if this is used somewhere.

2 Terminators are better

  • The theoretical reason: In separator notations there is one comma less than the number of elements. An empty list would need -1 commas, which can't be written, obviously. That is, empty lists must always be handled differently in the separator approach. There is no such problem with terminators.
  • The practical reason: In terminator notation, each list element is followed by the terminator symbol. Thus it is easier to reorder the elements of a list in an editor. If you have written
    (1:2:3:[])
    you can simply cut some elements and the subsequent ':' and then you can insert them whereever you want. For similar reasons terminator notation is easier to handle for VCS and diff.