An Information System (IS) can be viewed as managing a model of some slice of reality for the purpose of supporting inquiries. A variety of problems in information management, including database design, query optimization and seeking/integrating appropriate source of data, are known to benefit from knowledge concerning the semantics of the application problem. Research on knowledge representation and reasoning therefore has significant potential applications in IS. Description Logics (DLs) are a family of knowledge representation schemes that allow one to describe and define concepts describing objects in terms of their membership in classes and their inter-relationships to other objects. They have the advantage of well-defined semantics and well-studied properties concerning the complexity of deductions, including some efficient implementations. DLs have already been applied by researchers, including the author, to data management tasks such as intelligent data exploration and schema integration. This project investigates extensions of DLs that are motivated by the specific desire to support the development of better information systems. The extensions involve (1) making DLs more expressive as query and view definition languages by allowing variables to appear in descriptions for the purposes of pattern matching; (2) providing features for capturing uniqueness constraints, which express domain semantics very relevant to databases and query optimization; and (3) capturing significant aspects of Event-Condition-Action rules used in active databases, with the goal of supporting tools for reasoning about the consistency, interaction and inter-relationship of rules. The resulting languages and reasoners will support improved capture of data semantics, expressing more selective queries and presenting them more effectively, evaluating queries more efficiently, and the development of tools that help manage the DBMS components (schema, rules, etc.).
