A classification framework for pointcut languages in runtime monitoring

Abstract

In runtime monitoring and aspect-oriented programming, the execution of a program is monitored to check whether a property - formulated in a pointcut language - holds at some point in the execution of the program. Pointcut languages differ significantly in their expressiveness and the amount of information they utilize about the state of the program's execution, and the relation between different pointcut languages is often not clear. We propose a formal framework that provides the common abstractions of these languages and identifies the points of variability, such that pointcut languages can be compared and classified with regard to their expressiveness and cost. Besides its usage as a common frame of reference for pointcut languages, our framework also gives a precise model of the design space of pointcut languages and can hence help to design future pointcut languages in a more principled way. © 2009 Springer Berlin Heidelberg.