A type-theoretic interpretation of pointcuts and advice

Abstract

This article defines the semantics of MinAML, an idealized aspect-oriented programming language, by giving a type-directed translation from a user-friendly external language to a compact, well-defined core language. We argue that our framework is an effective way to give semantics to aspect-oriented programming languages in general because the translation eliminates shallow syntactic differences between related constructs and permits definition of an elegant and extensible core language.The core language extends the simply-typed lambda calculus with two central new abstractions: explicitly labeled program points and first-class advice. The labels serve both to trigger advice and to mark continuations that the advice may return to. These constructs are defined orthogonally to the other features of the language and we show that our abstractions can be used in both functional and object-oriented contexts. We prove Preservation and Progress lemmas for our core language and show that the translation from MinAML source into core is type-preserving. Together these two results imply that the source language is type safe. We also consider several extensions to our basic framework including a general mechanism for analyzing the current call stack.

Types and modularity for implicit invocation with implicit announcement

“…In [Ligatti et al 2006], Ligatti, Walker, and Zdancewic present formal semantics for an idealized AOPL. For this, they extend the simply-typed lambda calculus with two new abstractions covering join points, pointcuts, and advice, and prove type safety for this calculus.…”

“…MINIAML has some similarities with our language, most prominently that it allows scoping of advice: functions can be hidden from advice, thereby allowing "programmers to retain some control over basic information hiding and modularity principles in the presence of aspects." [Ligatti et al 2006] The mapping of MINIAML to the core calculus is non-trivial; we expect a corresponding mapping of our own language, although certainly desirable to prove the soundness of our type system, to be no easier. On the other hand, what we have deliv-ered can be immediately tried out in practical settings, allowing the community to test and improve it until it is maximally useful.…”

A Mechanized Model of the Theory of Objects

“…For example, Gordon and Hankin extended the ς-calculus towards the paradigm of parallel programming [15]. More recently, the ς-calculus has been incorporated into the ASP calculus that is a theoretical basis for distributed objects [11], and also into higher-level flavors like aspect-orientation [19].…”

Composing Safely — A Type System for Aspects

“…The basic idea of our calculus of aspects is similar to the theory of aspects [14] but we start from the Theory of Objects, unlike the former that is based on the λ-calculus. These models of aspects simply introduce labels in the base program.…”

“…Using this global assumption, we lose no generality. Ligatti et al [14] take a different line of argument to justify labels. They construct a core calculus that serves as the target code for a high-level aspect-oriented language.…”