[tome]

> A more serious problem with higher-kinded types IMO is that they wouldn't interact very well with an ML-style module system

That's interesting. Why is that?

[catnaroek]

Because, in ML, a type-level function that one module views as an abstract type constructor might be viewed by another module as a type synonym. Haskell's type language allows type constructors to appear partially applied, but requires type synonyms to appear fully applied, so it can't deal with this discrepancy.