Hello all!
There has been a discussion going on about if the list
constructor operators (:, [], and type []) should be dealt
with in the same way as with other function and type names.
I think there are two separate issues here:
1. Introduction of special syntactic identifiers for:
- the type constructor [], and the applied
version [a],
- the empty list constructor [] (and maybe
also the cons constructor :, but that one
fits in normal Haskell operator syntax).
2. Syntactic sugar which is translated away using prelude
functions. This includes:
- list notation, [1,2,3,4,5],
- list comprehensions, [ (x,y) | x <- xs, y <- ys ],
but also:
- do-notation,
- integer constants,
- (and maybe others).
About issue number 1, I don't think there is any problem in
making [] special syntax for some type constructor, of which
[a] is the applied version. This means [] is just a normal
(type constructor) name which we can import and export.
Issue number 2 is completely different and unrelated. Note
that this also includes normal prelude functions without
special syntax (such as >>=, return, mfail, fromInteger,
etc.). What happens in general when one uses this special
notation in a module which redefines these operators? I
think the easiest thing to do is just to make the
translation *always* refer to their prelude definitions.
Just my 2 �re...
Regards,
Koen.
--
Koen Claessen http://www.cs.chalmers.se/~koen
phone:+46-31-772 5424 e-mail:[EMAIL PROTECTED]
-----------------------------------------------------
Chalmers University of Technology, Gothenburg, Sweden
