Object-oriented modelling of periodicals

As new online systems to publish documents are quickly arising, more powerful models for multimedia documents are needed. To achieve new tools for efficiently managing these large data banks, it is not enough to represent the structure of documents in a standard format. To classify and represent properties of these publications are also important considerations. We apply an extended object oriented model to represent periodicals. The model focuses on the most advanced features of this kind of data: document structure of composition and temporal properties and behaviour. With so much information in a coherent model, we expect to contribute to the development of powerful new retrieval tools. The main advantages of our model are that: it reflects a document's identity at the same time as it defines various kinds of relationships between them; it provides flexibility in document structure and data types; and it allows one to consider documents at various levels of abstraction.