Defining Quantifiable Measures for Data Coupling and Control Coupling

We describe early results from a case study that applies a new approach to data coupling and control coupling (DC/CC) where coupling criteria can be expressed as quantifiable metrics based on multiple observation points in the code under test. The basis for the case study was a DO-178C DAL A display system application developed within Collins Aerospace. Examples of complex couplings are presented, and we report on how static analysis was used to build up a more complete criteria from the examples identified. An automated approach, based on static analysis and the dynamic execution of instrumented code during integration test provided the basis for proof-of-concept DC/CC coverage reporting. We also discuss how we envisage the approach fitting into a more general DC/CC framework that features two levels of rigor –one process-based and guideline-assisted, and the other based on static analysis to derive complex properties like those discovered in the study. Our conclusions revisit the importance of DC/CC analysis and highlight the benefits of a strong approach to it from a practitioner's perspective, especially with respect to how it supports design modularity.