Modeling with KRISYS: the Design Process of DB Applications Reviewed
Nelson Mendonça Mattos, Marjan Michels · 1989
During the last few years, various approaches to an extension of existing models with abstraction mechanisms have been proposed. Abstraction concepts enrich the semantics of the underlying data or knowledge model, but advocate modeling methodologies which are quite different from conventional ones: these mechanisms, in contrast to existing DB techniques, support a modeling process in a stepwise fashion. Our KBMS prototype KRISYS allows for such an incremental construction of data or knowledge bases by means of a complete and integrated support of the abstraction concepts provided by its knowledge model. In this paper, we show how the user can employ KRISYS to model some real world situations, thereby demonstrating that the abstraction concepts are at the focal point of the mentioned incremental modeling process. Since the early years of the Computer Science, most of the research activities in its various fields (Programming Languages, Software Engineering, Database Systems (DBS), Artificial Intelligence (AI), etc.) focuses on issues for facilitating the specification and management of the ever increasing complexity of software [1]. Based on the assertion that a computer system constitutes a model of the real world, about which it contains specific information pertinent to a particular application [2,3], the process of software development can be viewed as the process of building an accurate model of some enterprise. Thus, especially data modeling in the area of DBS and knowledge representation in the area of AI deal with the above mentioned software problem by working on the development of models that naturally and directly reflect the user's conceptualization of the real world [4]. This should facilitate the task of the modeler (as well as the interaction system/users), accelerating the process of software development as a whole. To achieve this goal, both fields try to improve the semantics of existing models by identifying and incorporating into them the constructs people apply during the process of describing the real world. In [4], we have shown that most of these constructs have their roots in epistemological methods for organizing knowledge; i.e., in trying to describe the real world, people usually involuntarily apply some