On Wed, Sep 24, 2008 at 09:08:49PM +0900, Benjamin L.Russell wrote:
I'm having difficulty in understanding the following behavior:
In GHCi:
Prelude> :type [] [] :: [a]
but:
Prelude> :type ([]) ([]) :: [a]
I.e., the types of both the empty-list '[]' and the one-tuple containing the empty-list '[]' are '[a]' (a list of a generic type variable).
According to "Chapter 2. Types and Functions" (see http://book.realworldhaskell.org/read/types-and-functions.html) of Real World Haskell (beta) (see http://book.realworldhaskell.org/beta/),
Haskell doesn't have a notion of a one-element tuple.
Why not? It seems that a tuple is similar to a list, except that the elements need not be all of the same type, and that a tuple, unlike a list, cannot be extended. Yet:
Note that it would be really annoying to have parentheses create a one-tuple. That would lead to such joyous fun as
1 + (2 + 3)
<interactive>:1:0: No instance for (Num (t)) arising from a use of `+' at <interactive>:1:0-8 Possible fix: add an instance declaration for (Num (t)) In the expression: 1 + (2 + 3) In the definition of `it': it = 1 + (2 + 3) (artist's conception). So some other syntax would be needed, but as Philippa points out, that would be a lot of syntactic heaviness for not much benefit. Python actually does have one-tuples: you create a one-tuple containing, say, the number 3 by writing: (3,) Ugly! In the end, there's no *theoretical* reason why Haskell doesn't have one-tuples: this is one of the few places in the language design where practical concerns won out over theoretical. -Brent