Abstract

The market forces affecting today's software development have placed a greater emphasis on software quality. A variety of design measurements have been proposed to control the development of software systems and to evaluate the final software products after delivery. In this thesis, we propose a slicing-based coupling measurement framework to assess the complexity of different granularities in a slice. The proposed framework combines well-known and proven coupling measurements, namely CBO, RFC, and MPC, with program slicing-based source code analysis techniques. These slicing-based coupling measurements are further measurements are implemented as part of the CONCEPT project to provide an aid for programmers comprehending and assessing source code and to give some objective heuristics for software engineers in selecting and prioritizing maintenance tasks.