In Search of a Unified Theory for Early Predictive Design Evaluation for Software

Traditional engineering design discipline calls for designs to be evaluated long before they are implemented. Early design evaluations predict properties of the artifact that will result from a proper implementation of the design and the value of those properties to the client or end user. The predicted properties can include costs as well as functionality, performance, and quality measures. Software engineering has some such evaluation techniques but the discipline lacks a systematic way to explain, compare, develop, and apply them. We discuss the role of early predictive design evaluation in software design, show how a variety of specific predictors serve this role, and propose a unifying framework, Predictive Analysis for Design (PAD) for design evaluation techniques. We are especially interested in techniques that predict the value of the finished software system to its client or end user and that make the predictions before the expense of software development or integration is incurred. We show that our PAD framework, even in its preliminary state, is sufficiently expressive to be useful in explaining and characterizing design evaluation techniques. We argue that the PAD framework shows sufficient promise to justify further development toward a unified theory of early predictive design evaluation.