Comprehensibility and efficiency of multiview framework for measurement plan design

摘要

Understanding the results of measurements is a primary issue for continuous software process improvement. Models provide support for better understanding measures. One of the problems often encountered in defining a measurement plan is its dimensions in terms of goals and metrics. This inevitably impacts on the usability of a measurement plan in terms of effort needed for interpreting the measurement results and accuracy of interpretation itself. The authors validate an approach (multiview framework) for designing a measurement plan, according to the GQM model, and structured in order to improve usability. For this reason an experiment was executed to validate the approach and provide evidence that a GQM designed according to the multiview framework is more usable, and that interpretation depends on the collected measures and is independent of who interprets them. In the experiment the authors verify that a measurement plan designed according to the proposed model doesn't negatively impact on efficiency of interpretation. The experimental results are positive and encourage further replications and studies.