Using data provenance to improve software process enactment, monitoring and analysis

A practice to support software processes continuous improvement is to reuse the knowledge acquired in previous executions. One way to capture process execution data is by using data provenance models. Data provenance refers to the origin, lineage or source of data. In computational terms, provenance is a historical record of the derivation of data that can help in its understanding. But, does the provenance data be used to contribute to the software process improvement? Based on this question, the approach proposed in this work aims to apply data provenance to support software process execution, monitoring and analysis phases as well as software process improvement as a whole. To achieve this goal, data provenance, ontology and predefined metrics were used in a pilot case study, considering software processes used in two real software development companies. With this study, two types of implicit information were derived: (1) information about artifacts that can increase new process instances runtime, and (2) information related to new agents that can be added to execute a task and contribute to the reduction of the task runtime.