Analyzing Data on Software Evolution Processes

This paper describes a tool supported process for evaluating software evolution processes using data analytics. Software evolution is concerned with correcting, changing and enhancing existing software systems. The evolution process is driven by error reports and change requests. The ISO-12207 standard prescribes how the process should be conducted. The approach presented here goes beyond simply checking if the process under investigation conforms to the prevailing standard. It collects data on the numbers and types of evolution requests as well as data on the time and effort required to implement them to determine if the evolution process is improving or degrading. For this, the error reports, change requests and time sheets are scanned to load a process performance database. There a numerical analysis is performed to measure the error and change rates for assessing process quality. Performance and usability data are also included in the quality assessment. In addition, the expended effort is analyzed in relation to the impact domain of the corrections and changes to determine the productivity rate. The number of reported errors and submitted change requests is essential to estimating the costs of future releases.