ITR: Controlling Software Complexity with Aspects and Analysis

Northeastern University, Boston MA

Investigators

Abstract

ABSTRACT CCR-0312598 PI: Mitchell Wand Northeastern University Modern software systems depend on the interaction of a large number of components. For such systems to function reliably, this complexity must be controlled. Traditional system-composition methods, such as modules and layering, have turned out to be brittle when confronted with the need for adaptation to new contexts. Aspect-oriented programming (AOP) is a promising recent technology for allowing adaptation of program units across module boundaries. However, AOP as currently practiced sacrifices many of the benefits of traditional module systems. The goal of this research is to develop a new language constructs for AOP that will retain the benefits of traditional module composition. A second goal is to extend program analysis, a well-known set of techniques for understanding programs in the small, to help understand the interaction of components in AOP systems. Such analysis techniques also have implications for reliable scientific computing.