TESTING AND TESTABILITY OF OBJECT-ORIENTED SOFTWARE SYSTEMS VIA METRICS

This dissertation investigates the factors that affect testability and testing cost of object- oriented software systems. Developing a software program which eases the testing process by increasing testability is crucial. Also, to assess whether or not the testing effort and cost consumed or planned is adequate or not is another critical matter this dissertation aims to answer by composing a new way to evaluate the links between software design parameters and testing effort via source-based metrics. An automated metric plug-in is used as the primary tool for obtaining the metric measurements. Our study is based on the investigation of many open-source projects written in Java to achieve our goals. By the help of the statistical evaluation of project data, we both propose a new model to assess testing effort and testability, and find significant relations and associations between software design and testing effort and testability of object- oriented software systems via source-based metrics.