Monitoring and Improving the Quality of ODC Data using the "ODC Harmony Matrices": A Case Study

Abstract

Orthogonal defect classification (ODC) is an advanced software engineering technique to provide in-process feedback to developers and testers using defect data. ODC institutionalization in a large organization involves some challenging roadblocks such as the poor quality of the collected data leading to wrong analysis. In this paper, we have proposed a technique ('Harmony Matrix') to improve the data collection process. The ODC Harmony Matrix has useful applications. At the individual defect level, results can be used to raise alerts to practitioners at the point of data collection if a low probability combination is chosen. At the higher level, the ODC Harmony Matrix helps in monitoring the quality of the collected ODC data. The ODC Harmony Matrix complements other approaches to monitor and enhances the ODC data collection process and helps in successful ODC institutionalization, ultimately improving both the product and the process. The paper also describes precautions to take while using this approach