ONTOLOGY FOR SOFTWARE METRICS AND INDICATORS

Abstract

Software and even more web measurement -as a younger discipline, are currently in a stage in which terminologies, models, and methods are still being defined and consolidated. It is a necessity to start reaching a common agreement between researchers and other stakeholders about primitive concepts such as attribute, metric, measure, measurement and calculation method, scale, elementary and global indicator, calculable concept, among others. There are various useful recently issued ISO standards related to software quality models, measurement, and evaluation processes; however, we observe sometimes a lack of a sound consensus among the same terms in different documents or, sometimes, absent terms. In this manuscript, we present an ontology for software metrics and indicators -based as much as possible on the concepts of those standards, which can be useful to support different assurance processes, methods and tools, in addition to be the foundation for our cataloging web system. In order to illustrate the ontology, we focus particularly on a set of intermediate representations for the domain (such as UML diagrams and tables), which were yielded during the conceptualisation step. In addition, a discussion about decisions that have been taken in choosing the terms is presented. Without sound and consensuated definition of terms, attributes, and relationships it is difficult to assure metadata consistency and, ultimately, data values are comparable on the same basis.