Investigating the relation between static and dynamic coupling metrics and the fault-proneness of object-oriented software systems
Atam Gangaram Panday
The problem definition of this Thesis is to examine the predictability of the fault-proneness of object-oriented software of dynamic coupling metrics versus static coupling metrics, not disregarding the fact that object-oriented software systems may differ in their level of usage of object-oriented features, which may have an impact on the quality of dynamic measures as an indicator.
The main contribution in this thesis is stating that object-oriented software differs in their usage of object-oriented features which is the basis for object-orientated metrics, especially dynamic ones. This causes differences when analyzing dynamic metrics of different object-oriented software systems. To help dealing with this phenomenon a new metric DII, dynamic invocation indicator, is defined. Besides DII, a few dynamic coupling metrics are defined along with a new coupling type: Neighbor Coupling, which is also considered in the analysis carried out. A tool “Jrev” was developed for extracting al these metrics along with some static coupling metrics and some size metrics from an input Java software system. This tool will be made available for others. The final contribution of this thesis project is supporting the usage of a framework in analyzing metrics data, which will eventually (hopefully) lead to some standard in software metrics analysis in order for research in this area being more consistent and easier for interpretation.
Two Java software systems, Velocity and Tomcat, and investigated in this thesis.
Main.AtamGangaramPanday moved from Main.AtamPanday on 15 Oct 2006 - 19:39 by ArieVanDeursen
- put it back