Using Functional Defect Analysis as an Input for Software Process Improvement: Initial Results

In this paper we present how functional defect analysis can be applied for software process improvement (SPI) purposes. Software defect data is shown to be one of the most important available management information sources for SPI decisions. Our preliminary analysis with three software companies’ defect data (11653 defects in total) showed that 65% of all the defects are functional defects. To better understand this mass, we have developed a detailed scheme for functional defect classification. Applying our scheme, defects can be classified with accuracy needed to generate practical results. The presented scheme is at initial stages of validation and has been tested with one software company’s defect data consisting of 1740 functional defects. Based on the classification we were able to provide the case organization with practical improvement suggestions.