Modelling and measurement of Prolog data

In software technology, the need for measurement is well recognised, but it is proving difficult to devise measures that are unique, meaningful, objective and representative. A method is given for characterising inherent properties of static Prolog code and assigning to them values by measurement. The principal aim is to demonstrate that a model-based approach to measurement helps to ensure the quality of measures. The authors propose a model of the data entities in Prolog program texts, together with the model of the constructors that combine them into complete programs. The model-based measurement system presented characterises data entities by a six-tuple, and composes six-tuples to yield measures of complete programs or their parts. Model-based measurement ensures that the meaning of the measures is explicit