No. He does not mean unifying separate languages or dialects into one standard.
Rather, the point is that the existing ML language has two parts that are pretty unrelated: the normal value-level language where you write your actual programs, and the module-level language you use for organizing your code and types.
Think of it like the distinction between C proper and the C preprocessor. They're part and parcel of the same language in practical terms, but really they're two completely unrelated languages grafted into one. The ML situation isn't nearly as bad, of course, but the idea is the same: wouldn't it be nice to have one language for both?
There are some ad-hoc efforts to bridge the gap between the two (extensions for first-class modules), but they are not enough. His solution is a language design that starts from first-class modules and, roughly, ensures the module language is just a special application of the base language. It radically simplifies the design and exposes the minimal underlying abstraction which is a typed variant of the λ-calculus called Fω.
Rather, the point is that the existing ML language has two parts that are pretty unrelated: the normal value-level language where you write your actual programs, and the module-level language you use for organizing your code and types.
Think of it like the distinction between C proper and the C preprocessor. They're part and parcel of the same language in practical terms, but really they're two completely unrelated languages grafted into one. The ML situation isn't nearly as bad, of course, but the idea is the same: wouldn't it be nice to have one language for both?
There are some ad-hoc efforts to bridge the gap between the two (extensions for first-class modules), but they are not enough. His solution is a language design that starts from first-class modules and, roughly, ensures the module language is just a special application of the base language. It radically simplifies the design and exposes the minimal underlying abstraction which is a typed variant of the λ-calculus called Fω.