Validation in the Software Metric Development Process

In this paper the validation of software metrics will be examined. Two approaches will be combined: representational measurement theory and a validation network scheme. The development process of a software metric will be described, together with validities for the three phases of the metric development process. Representation axioms from measurement theory are used both for the formal and empirical validation. The differentiation of validities according to these phases unifies several validation approaches found in the software metric's literature