CBR for Experimental Software Engineering

Abstract

The objective of our work is to exploit the mutual interrelations between case-based reasoning and experimental software engineering (ESE) for the sake of both fields. In particular, we address the following topics:

• Presentation of a logical infrastructure of organizational learning in the software domain that makes use of principles, methods, and established practices from CBR.

• Utilization of CBR technology for implementing experience bases.

• Outlining of additional uses of CBR for ESE that go beyond infrastructures for organizational learning (i.e., data analysis, problem solving, etc.).

• Evolution of a methodology for developing CBR-based software systems based on software engineering methods and practices that are established in other application domains.

These topics are gradually developed throughout this chapter. We start with brief introductions to software engineering and ESE (Sections 9.2 and 9.3). Following this, we outline possible CBR support for ESE and give an overview of research related to both CBR and ESE (Sections 9.4 and 9.5). In the main part of this chapter, we describe the approach we take at the Fraunhofer Institute for Experimental Software Engineering in Kaiserslautern (Section 9.6). We also report on two case studies and give an outlook on work still to be done (Sections 9.7 and 9.8).