Software undergoes changes at all stages of the software development process. Accepting too many changes will cause expense and delay and rejecting the changes may cause customer dissatisfaction. One of the inputs that help the software project management to decide whether to accept or reject the changes is by having reliable predictions of the impact of the changes. Change impact analysis is one of the methods that can be used to provide the predictive information. Many current impact analysis techniques have been developed for the software maintenance phase. These techniques assume that all classes in the class artifact are completely developed and the class artifact is used as a source of analysis since it represents the final user requirements. However, these assumptions are not practical for impact analysis in the software development phase as some classes in the class artifact are still under development or partially developed. This leads to inaccuracy. This thesis presents a novel impact analysis framework to be used in the software development phase. The framework composes two main stages that are the development of the class interactions prediction and the impact analysis. The significant achievements of the framework are demonstrated through an extensive experimental validation using several case studies. The experimental analysis shows improvement in the accuracy over current impact analysis results.
|Qualification||Doctor of Philosophy|
|Publication status||Unpublished - 2011|