More Modular Reasoning for Aspect-Oriented Programs

Iowa State University, Ames IA

Investigators

Abstract

Proposal 0428078 TITLE More Modular Reasoning for Aspect-Oriented Programs Gary T. Leavens This project will advance the theory and practice of aspect-oriented software development. Aspect-oriented languages allow modularization of code for requirements that do not align with a program's architecture, reducing scattering and tangling of code for these cross-cutting requirements. The problem addressed by this project is how to recover static, modular reasoning in aspect-oriented programs. The technical approach is to create the design discipline, based on an analogy to behavioral subtyping, and a small aspect-oriented programming language to support it. By producing prototype tools and applying them in case studies, the researchers will experimentally validate the utility of the language and design discipline. A formal study of specification and verification techniques will demonstrate soundness of static, modular reasoning in the language and design discipline. The discipline developed through the project will benefit software engineers and programmers by providing guidance for thinking about and applying aspect-oriented programming techniques. The discipline may also improve software correctness and maintainability. By showing how a language can support the discipline, the project will contribute to the design of programming and specification languages. Like the notion of behavioral subtyping, a disciplined approach to thinking about aspect-oriented abstraction will help educators convey the key ideas of the paradigm.