Incomplete objects-a data model for design and planning applications

Abstract

We are motivated by applications within design, planning and scheduling areas, where current research appears to be focused on syntactic issues of performance and volume. We take a more semantic view of applications within these areas and discover several useful functionalities that are poorly supported. For example, facilities for handling incomplete specifications are quite inadequate. We introduce a notion of OR-objects and show that it captures incomplete specifications naturally. In particular, a database with OR-objects represents a set of possible worlds, e.g., a world for each design or schedule, and queries can either be evaluated in the "interpretations" of the database, or in the database itself. We formalize these notions of interpretations and hypothetical queries in an object-oriented setting, and provide a complexity characterization for our queries.