Re: [Haskell-cafe] looking for examples of non-full Functional Dependencies

I think I was the one confused. I guess I was (falsely) thinking that both C Int Char T C Char Int T could both be instances of class C a b c | c -> a, c -> b but only one could be an instance of C a b c | c -> a b. Sorry for adding noise to the discussion. Ryan Ingram wrote:

I'm still confused about this point:

On 4/16/08, Dan Weston wrote:

...

...

...

class C a b c | c -> a b

Notice that there are multiple (two) parameters in the range of the FD.

It's tempting to convert the above to

class C a b c | c -> a, c -> b

but this yields a weaker (in terms of type improvement) system.

In both cases the statement is that given a type x, the instance C x y z for some y,z and the constraint C x a b, we unambiguously have a ~ y, b ~ z (where ~ is type equality)

How does the order in (c -> a b) matter?

-- ryan