Talk on "Insight on Role Modeling"

by Amandio Vaz Velho, INESC

Real-world entities are intrinsically evolutionary and multi-perspectived. To capture those characteristics in an abstract way, data models should include role modeling features. In spite of its significance, role modeling has been neglected by the data model community. Fortunately, recently, this picture has been progressively being changed. Nevertheless, there isn't yet a comprehensive understanding of role modeling issues.

This talk presents a survey of role modeling. A role is defined as a structure, behavior and visualization unit which can be dynamically acquired and relinquished. A taxonomy - conceptual taxonomy - is introduced which states characteristics of real-word entities that role models should address. The reasons why most models currently available are unable to capture those issues are clarified. A second taxonomy - implementation taxonomy - is introduced that synthesizes alternative programming mechanisms. These mechanisms can be incorporated in the models to support the dimensions of the conceptual taxonomy. Finally, a set of role models recently proposed are surveyed. They are classified according to both taxonomies. The talk provides extensive biography on the most relevant role modeling work.