Jason,

Oh sorry. I definitely meant not available. I hope it was available though :)
I'll look into dependent types, sure. But as I said, I can come up with a way (even though it is ugly) via promoting every constructor to a data type and using type classes to group different types. This makes me think there can be a more elegant way of handling it.

Matthias,

I searched in the list, but couldn't find the discussion you've reffered. Do you have a link to it by any chance?


Thanks for the comments. Anyone else?


On 19 March 2010 19:39, Jason Dagit <dagit@codersbase.com> wrote:


On Fri, Mar 19, 2010 at 8:15 AM, Ozgur Akgun <ozgurakgun@gmail.com> wrote:
Hi Cafe!

Disclaimer: I know what I'm going to ask is now available as a language feature normally.

Did you mean "not available"?  I don't know of a Haskell language feature for this, so if you really did mean "now" then I'd love to learn about the feature you're thinking of.

Only think I can think of is having the constructors, as seperate data types, introducing new type classes to group every possible subset of [X,Y,Z] and [A,B,C,D] and use those type classes when defining the functions. But as you can imagine, this is not the only place I want to use this thing, and the code will start to look cryptic if I do so. I must add, I want to use this disjunction of constructors thingy in data type declerations as well.

It looks to me like you want dependent types.  You might look at Agda.  It's similar to Haskell but features dependent types and other interesting things.

Jason



--
Ozgur Akgun