Ontologies can be viewed as the middle layer between pure human conceptualization and machine readability. However, they have not lived up to their promises so far. Most ontologies are too tailored to specific data and use-cases. By making sometimes strong, or sometimes too weak, ontological commitments, many existing ontologies do not adequatly reflect human conceptualizations. As a result, sharing and reuse of ontologies is greatly inhibited. In order to more effectively preserve this notion of human conceptualization, an ontology should be designed with modularity and extensibility in mind. A modular ontology thus may act as a bridge between human conceptualization and data.
展开▼