[Haskell-cafe] 'type' for isomorphism [was poorly named: aggressiveness of functional dependencies]

[Bulat Ziganshin]

Hello Nicolas,

Friday, November 17, 2006, 1:17:30 AM, you wrote:

> So it works, but it wasn't quite what I had hoped for. Has there been
> any work/proposals for specifying a "finalized" type class? I.E. Those
> not subject to overlap and thus "closed" to the world?

it is a common idea, at least i had it and seen John Meacham's proposal

reasons mainly falls into two categories:

1) optimization: this allows compiler to make final decisions and
inline appropriate code instead of be ready for some new instances
that will change the whole game

2) with closed class, compiler may be smarter about overlapping, type
deriving and other type checking/inferring procedures


-- 
Best regards,
 Bulat                            mailto:Bulat.Ziganshin at gmail.com