Uma Visão Ontológica de Aspectos Comportamentais e Estruturais da Medição

Abstract

Measurement is a key process in several domains. It is able to assign values to properties of entities, quantifying them and thus supporting the generation of conclusions and decision making. It is, perhaps, one of the most fundamental concepts present in science. Without the ability to measure, it would be impossible for scientists to conduct their experiments or form their theories. Although measurement has particularities in each application domain, part of the measurement knowledge crosses the domain boundaries and is common to several of them. However, there is no single and complete definition of the measurement process, nor of the concepts involved in this process. When there is no common conceptualization, there is often divergence among norms, standards and other process models that address measurement. Considering this scenario, this work proposes a task ontology, seeking a common and shared definition about measurement, which deals with both behavioral (e.g., flows among activities) and structural (e.g., objects manipulated by activities) aspects. The Measurement Task Ontology (MTO) proposed in this work extends the previous version of MTO, by making it capable of covering new activities and concepts, and more properly addressing the process dynamics. MTO is composed of a structural model, which itself represents a Core Ontology on Measurement (COM). COM was also extended in this work to cover the concepts and activities inserted into MTO. In order to achieve conceptual clarity and fidelity to reality, the ontologies were grounded on the Unified Foundational Ontology (UFO). Considering that core ontologies have high potential to reuse, to facilitate the reuse of the conceptualization provided by COM in different application domains, it was organized as an ontology pattern language called Measurement Ontology Pattern Language (M-OPL), which guides the ontology engineer in applying the patterns needed for modeling the domain of interest. To support the use of M-OPL, M-OPL Tool, a computational tool, was developed and used in a study to evaluate the use of M-OPL to support the development of ontologies for specific domains.