Abstract

Accurate and expressive representation of the subject matter over which a context-oriented, decision-support system operates is fundamental to the effectiveness and longevity of the resulting solution. Often taking the form of an ontology, such extensive representational models, by their very nature, are rich in relationships and both coarse and fine-grained objects. It is, however, these qualities enabling rich expression that can significantly increase both the complexity of developing against these models as well as the potential for incurring undesirable performance issues. Further, due to the typically detail-oriented usage inherent in the software-based users (i.e., reasoning agents, etc.) of these models, it is important to recognize that a singular view of the world so to speak is not necessarily appropriate across the entire Ontology user base. In fact, in such highly expressive environments, it is critical to not only recognizing these distinctions in user perspective, but to, in fact, promote and exploit them. It is by acknowledging and consequentially supporting this perspective-based individuality among Ontology users that true representational accuracy and utility is achieved.

Traditionally, software-based users comprising decision-support systems have operated over a singular, common representation. However, in the Perspective Model-enriched environment presented in this paper1, Ontology users are empowered with the ability to effectively perceive the world in accordance with individualized, native views. These views are then seamlessly inter-linked with one another to form a multi-Perspective Model of the target domain capable of supporting rich interoperability. Exclusively operating over personalized Perspective Models, users are not only shielded from the broad-scoped complexities inherent in the more omniscient concerns of the Ontology’s entire scope but are also able to both view and interact with it in terms of more native representation.

To be effective, the concept of Perspective Models must be partnered with a supportive model development process. In addition to an explanation of the concept of Perspective Models, this paper also presents a purpose-built development process that supports effective creation of the potentially numerous sets of models inherent in this type of expressive paradigm. The process offered in this paper effectively parcels the development of individual Perspective Models with the individuals possessing the necessary domain and use-case expertise. In this manner, the development process strives to significantly increase the involvement of the entire set of team members in the modeling activity, both capitalizing on user domain expertise in addition to increasing critical user understanding as well as acceptance of the representation over which their components will operate.

Disciplines

Software Engineering

 

URL: http://digitalcommons.calpoly.edu/cadrc/65