Dependence clusters in source code
ACM Transactions on Programming Languages and Systems (TOPLAS)
| Hi-index | 0.00 |
In this paper we describe our on-going longitudinal study of a large complex software development project. We discuss how we used project metrics data collected by the development team to identify threats to project outcomes. Identifying and addressing threats to projects early in the development process should significantly reduce the chances of project failure. We have analysed project data to pinpoint the sources of threats to the project. The data we have used is embedded in the projectýs fortnightly progress reports produced by the project team. The progress reports are part of the software measurement program this company operates. The company has highly mature development processes which were assessed at CMM level 5 in 2004. Our analysis shows that standard project progress data can generate rich insights into the project; insights that go beyond those anticipated when the metrics were originally specified. Our results reveal a pattern of threats to the project that the project team can focus on mitigating. The project team is already aware of some threats, for example that communication with the customer is a significant threat to the project. But there are other threats the team is not aware of, for example that people issues within the software team are not a significant threat to the project.