Software Development Process Monitoring Based on Nominal Transformation

For any software project, a development plan is needed to distribute project tasks to different teams and to manage team activities. For development process, the attribute value of the process to complete a task will vary around nominal value. In general, when the value exceeds a certain amount of specification limits, it is thought that the software process has some problems and intervention measures are needed. But in statistics, the process might be running well and any intervention will cause chaos. In order to monitor the team software process, it is needed to build up a rule to indicate if the process is running stable or not. In this paper, a nominal transformation method for measured task data based on task expectation is introduced to unify the measured data from different processes. The statistical characteristics for the measured data are analyzed. The control charts are given based on the statistical control theory and an example is presented to show the way that monitors software development process.