Section: New Results

Aspect-Oriented Programming

Participants : Rémi Douence, Guilhem Jaber, Ismael Mejía, Jacques Noyé, Mario Südholt, Nicolas Tabareau.

We have contributed to the foundations of aspect-oriented programming and presented new programming languages for aspects and related paradigms.

Formal models for AOP

We have presented two calculi contributing to the foundations of AOP: the A Calculus, a parameterized calculus encompassing AspectJ-like and history based aspect languages, and a category-theoretic definition of AOP in terms of 2-categories.

The A Calculus

With partners from Vrije Universiteit Brussel and Aarhus University, we have extended the foundational calculus for AOP (introduced in 2010) that supports the most general aspect model to-date compared to existing calculi and the deepest integration with plain OO concepts [12] . Integration with OOP is achieved essentially by modeling proceed using first-class closures. Two well-known pointcut categories, call and execution that are commonly considered similar are shown to be significantly different; our calculus enables the resolution of the associated soundness problems. The A-calculus also includes type ranges, an intuitive and concise alternative to explicit type variables that allows advices to be polymorphic over intercepted methods. Finally, our calculus is the first aspect calculus to use calculus parameters to cover type safety for a wide design space of other features. The soundness of the resulting type system has been verified in Coq.

In 2012, we have covered a range of choices with respect to evaluation order and non-determinism. We have studied one version that enforces a deterministic call-by-value semantics, and another one that omits restrictions on evaluation order and allows many kinds of non-determinism. Furthermore, we have provided a mechanized complete type soundness proof using the theorem prover Coq.

A category-theoretic foundation of aspects

Aspect-Oriented Programming (AOP) started fifteen years ago with the remark that modularization of so-called crosscutting functionalities is a fundamental problem for the engineering of large-scale applications. However, theoretical foundations of AOP have been much less studied than its applicability.

We have proposed [26] to put a bridge between AOP and the notion of 2-category to enhance the conceptual understanding of AOP. Starting from the connection between the $\lambda$-calculus and the theory of categories, we have defined an internal language for 2-categories and shown how it can be used to define the first categorical semantics for a realistic functional AOP language. We have then used this categorical framework to introduce the notion of computational 2-monads for AOP. We have illustrated their conceptual power by defining a 2-monad for Éric Tanter's execution levels—which constitutes the first algebraic semantics for execution levels—and then introducing the first exception monad transformer specific to AOP that gives rise to a non-flat semantics for exceptions by taking levels into account.

Programming languages for aspects and related paradigms

We have introduced three results related to aspect-based programming languages: an extension of EScala for multi-paradigm programming; Monascheme, a language for modular prototyping of aspect-based languages and language support for membranes, an aspect-based means for structuring computations.

Concurrent multi-paradigm programming with EScala

EScala integrates, around the notion of declarative events, object-oriented, aspect-oriented and event-based programming [30] . However, in spite of the fact that events naturally evoke some form of concurrency, there is no specific support for concurrency in EScala. It is up to the programmer to understand how to combine declarative events and Scala's support for concurrent programming. We have started working on injecting concurrency at the heart of declarative events so that events can indeed be naturally concurrent [28] .

Monascheme: modular prototyping of aspect languages

We have then developed Monascheme [21] , an extensible aspect-oriented programming language based on monadic aspect weaving. Extensions to the aspect language are defined as monads, enabling easy, simple and modular prototyping. The language is implemented as an embedded language in Racket. We illustrate the approach with an execution level monad and a level-aware exception transformer. Semantic variations can be obtained through monad combinations.

Structuring computations with aspect-based membranes

In most aspect-oriented languages, aspects have an unrestricted global view of computation. Several approaches for aspect scoping and more strongly encapsulated modules have been formulated to restrict the power of aspects. Our approach [27] leverages the concept of programmable membranes of Boudol, Schmitt and Stefani, as a means to tame aspects by customizing the semantics of aspect weaving locally. Membranes have the potential to subsume previous proposals in a uniform framework. Because membranes give structure to computation, they enable flexible scoping of aspects; because they are programmable, they enable visibility and safety constraints, both for the advised program and for the aspects. The power and simplicity of membranes open interesting perspectives to unify multiple approaches that tackle the unrestricted power of aspects.