[Haskell-cafe] existentially quantified data types - restrictions

Ozgur Akgun ozgurakgun at gmail.com
Thu Mar 25 11:07:40 EDT 2010

Dear Cafe,

I need to use a language feature which is explicitly documented to be a
restriction, and -even worse- I think I reasonably need to use it.

f2 (Baz1 a b) (Baz1 p q) = a==q
> It's ok to say a==b or p==q, but a==q is wrong because it equates the two
> distinct types arising from the two Baz1 constructors.
> [from * Restrictions* at
> http://www.haskell.org/ghc/docs/latest/html/users_guide/data-type-extensions.html
> ]

To simplify, let's say Baz is the only constructor of a data type,

data Baz = forall a. Eq a => Baz a

-- | this cannot be done:
instance Eq (Baz a) where
    (Baz x) == (Baz y) = x == y

I am quite tempted to use show functions for this equality comparison, but
after trying to have a nicely type framework I really don't want to do that.
What I simply want is, haskell to be able to compare them if they belong to
the same type, and return False otherwise. (not that haskelly way of doing
things, I know.)

Any suggestions better than the following are very welcome:
    (==) = (==) `on` show


Ozgur Akgun
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.haskell.org/pipermail/haskell-cafe/attachments/20100325/43f56cdb/attachment.html

More information about the Haskell-Cafe mailing list