Abstract

Two of the distinguishing features of Standard ML Modules are its term dependent type syntax and the use of type generativity in its static semantics. From a type-theoretic perspective, the former suggests that the language involves first-order dependent types, while the latter has been regarded as an extra-logical device that bears no direct relation to typetheoretic constructs. We reformulate the existing semantics of Modules to reveal a purely second-order type theory. In particular, we show that generativity corresponds precisely to existential quantification over types and that the remainder of the Modules type structure is based exclusively on the second-order notions of type parameterisation, universal type quantification and subtyping. Our account is more direct than others and has been shown to scale naturally to both higher-order and first-class modules.

Citations

573	The Definition of Standard ML (Revised - Milner, Tofte, et al. - 1997
251	A type-theoretic approach to higher-order modules - Harper, Lillibridge - 1994
208	Manifest types, modules, and separate compilation - Leroy - 1994
130	Higher-order modules and the phase distinction - Harper, Mitchell, et al. - 1990
118	Using dependent types to express modular structure - MacQueen - 1986
98	On the type structure of Standard ML - Harper, Mitchell - 1993
96	Applicative functors and fully transparent higher-order modules - Leroy - 1995
54	Types for Modules - Russo
53	A syntactic theory of type generativity and sharing - Leroy - 1996
53	Translucent sums: A foundation for higher-order module systems - Lillibridge - 1997
37	An interpretation of Standard ML in type theory - Harper, Stone - 1997
13	Pure Type Sytems formalized - McKinna, Pollack - 1993