In recent years, this distinction between the requirements and problems in KR and DB is vanishing rapidly. On the one hand, a modern KR-system must be able to handle large data sets if it is to be employed in realistic applications. This means that techniques developed in the DB area can and should be employed. On the other hand, the information stored in DBs becomes more complex, and thus requires more intelligent retrieval and reasoning techniques. For example, it turned out that important information about the connection between different data items could not be expressed in traditional data models. This led to the introduction of semantic, deductive, and object oriented data models. Recently, it has been shown that many of these data models can be expressed in suitable KR formalisms, which allows one to apply reasoning techniques from AI to database problems.
Unlike its predecessor workshops KRDB'94 and KRDB'95, which concentrated on the connection between object-oriented formalisms in KR and DB, KRDB'96 is intended to have a broader scope. We want to bring together researchers and developers from all areas of KR and DB where an interaction seems to be promising. In addition, users from industry can obtain a good impression of the research done on the border line between the two areas, and they can contribute their knowledge of what type of research is relevant in their applications.