The contribution of the paper is twofold. First, we define a general notion of type system equipped with an entailment relation between type environments; this generalisation serves as a pattern for instantiating type systems able to support separate compilation and interchecking of Java-like languages, and allows a formal definition of soundess and completeness of inter-checking w.r.t. global compilation. These properties are important in practice since they allow selective recompilation. In particular, we show that they are guaranteed when the type system has principal typings and provides sound and complete entailment relation between type environments.

The operation of expansion on typings was introduced at the end of the 1970s by Coppo, Dezani, and Venneri for reasoning about the possible typings of a term when using intersection types. Until recently, it has remained somewhat mysterious and unfamiliar, even though it is essential for carrying out compositional type inference. The fundamental idea of expansion is to be able to calculate the effect on the final judgement of a typing derivation of inserting a use of the intersection-introduction typing rule at some (possibly deeply nested) position, without actually needing to build the new derivation.

System E is a recently designed type system for the #- calculus with intersection types and expansion variables. During automatic type inference, expansion variables allow postponing decisions about which non-syntax-driven typing rules to use until the right information is available and allow implementing the choices via substitution.

We propose a rank 2 intersection type system for a language of modules built on a core ML-like language. The principal typing property of the rank 2 intersection type system for the core language plays a crucial role in the design of the type system for the module language. We first consider a "plain" notion of module, where a module is just a set of mutually recursive top-level definitions, and illustrate the notions of: module intrachecking (each module is typechecked in isolation and its interface, which is the set of typings of the defined identifiers, is inferred); interface interchecking (when linking modules, typechecking is done just by looking at the interfaces); interface specialization (interface intrachecking may require to specialize the typing listed in the interfaces); principal interfaces (the principal typing property for the type system of modules); and separate typechecking (looking at the code of the modules does not provide more type information than looking at their interfaces). Then we illustrate some limitations of the "plain" framework and extend the module language and the type system in order to overcome these limitations. The decidability of the system is shown by providing algorithms for the fundamental operations involved in module intrachecking and interface interchecking.

Useful type inference must be faster than normalization. Otherwise, you could check safety conditions by running the program. We analyze the relationship between bounds on normalization and type inference. We show how the success of type inference is fundamentally related to the amnesia of the type system: the nonlinearity by which all instances of a variable are constrained to have the same type.

Let # be a rank-2 intersection type system. We say that a term is #-simple (or just simple when the system # is clear from the context) if system # can prove that it has a simple type. In this paper we propose new typing rules and algorithms that are able to type recursive definitions that are not simple. At the best of our knowledge, previous algorithms for typing recursive definitions in the presence of rank-2 intersection types allow only simple recursive definitions to be typed. The proposed rules are also able to type interesting examples of polymorphic recursion (i.e., recursive definitions rec {x = e} where di#erent occurrences of x in e are used with di#erent types). Moreover, the underlying techniques do not depend on particulars of rank-2 intersection, so they can be applied to other type systems.

We propose a rank 2 intersection type system for a language of modules built on a core ML-like language. The principal typing property of the rank 2 intersection type system for the core language plays a crucial role in the design of the type system for the module language. We first consider a "plain" notion of module, where a module is just a set of mutually recursive top-level definitions, and illustrate the notions of: module intrachecking (each module is typechecked in isolation and its interface, which is the set of typings of the defined identifiers, is inferred); interface interchecking (when linking modules, typechecking is done just by looking at the interfaces); interface specialization (interface intrachecking may require to specialize the typing listed in the interfaces); principal interfaces (the principal typing property for the type system of modules); and separate typechecking (looking at the code of the modules does not provide more type information than looking at their interfaces). Then we illustrate some limitations of the "plain" framework and extend the module language and the type system in order to overcome these limitations. The decidability of the system is shown by providing algorithms for the fundamental operations involved in module intrachecking and interface interchecking.

this paper we show how they can be recast into algorithmic inference rules by applying a combination of standard techniques, among which constraint handling plays a central role. We then analyse the generated constraint sets and show that they share a common structure which can be simplified with a specialised algorithm we present. As the constraint simplification procedure is mainly based on unification, we implemented the whole inference algorithm in a Prolog program which we will comment upon