[Haskell] Plea for separate kind annotations

Just quickly, in light of some wiki tracs on Haskell prime i've read recently. I wonder if any one else is also of the opinion that kind annotations should be available and that they should be separated from type definitions in the same way as type annotations are separated from value definitions. Example: data T :: * -> * data T a = T a There are two levels of lambda calculus (values and types), with types structuring values and kinds structuring types. So why not let programmers deal with them in similar ways?