Name: ROBERTO CARRARETTO
Publication date: 28/11/2012
Advisor:
Name |
Role |
|---|---|
| JOÃO PAULO ANDRADE ALMEIDA | Advisor * |
Examining board:
| Name |
Role |
|---|---|
| GIANCARLO GUIZZARDI | Internal Examiner * |
| JOÃO PAULO ANDRADE ALMEIDA | Advisor * |
Summary: Many authors have discussed the importance of ontological concerns in the development of information systems, emphasizing the benefits of ontology-based approaches to conceptual modeling tasks. A principled ontology-driven approach typically relies on the definition of a domain ontology and its use in subsequent phases of information system design and integration. Many of the
challenges in the application of such an approach are related to addressing ontological concerns (defining the nature of phenomena of interest) and addressing informational concerns (defining the information demand about the phenomena of interest). In this thesis, we argue that ontological concerns should be clearly separated from informational concerns. We have observed that ontologybased approaches and information modeling approaches have been treated mostly in isolation, with the consequence that the relation between a domain ontology and an information model is still in need of clarification, despite the efforts of the formal ontology and the information modeling communities.
Therefore, in this thesis, we analyze conceptual modeling in terms of two levels, namely, the so-called ontological level and, what we call, the information level. Our initial effort is to characterize the information level in harmony with existing works on information modeling and to align our information level approach with the existing ones concerning the ontological level. Then, we strive to provide a model-driven approach in which a domain ontology addressing ontological concerns (at the ontological level) is used as a starting point for the definition of an information model addressing informational concerns (at the information level). Our model-driven approach is guided by several systematic informational decisions that we identify here, which assist the addressing of an information demand. We adopt a philosophically well-founded profile of the Unified Modeling Language (UML) class diagrams, called OntoUML, to represent domain ontologies. Further, we adopt UML class diagrams to represent information models in an object-oriented approach. Finally, we provide tool support for the model transformation from OntoUML to UML in order to operationalize the approach and show its technical feasibility.
