A Constraint-based Object Model for Structured Document Management From: ftp://ftp.eecs.umich.edu/people/rundenst/papers/r-95-6.data TITLE: A Constraint-based Object Model for Structured Document Management AUTHORS: Anisoara Nica and Elke Angelika Rundensteiner ABSTRACT Complex multimedia document handling, including the modeling, decomposition, and search across digital documents, is one of the primary services that must be provided by digital library systems. In this paper, we present a general approach for handling structured documents (e.g., SGML documents) by exploiting object-oriented database technology. For this purpose, we propose a constraint-based object model capable of capturing in a uniform manner all SGML constructs typically used to encode the structural organization of complex documents. We present a general strategy for mapping arbitrary document types (e.g., article, journal, and book DTDs) expressed using standard SGML into our model. Most importantly, we demonstrate that our model is designed to handle the integration of diverse document types into one integrated schema, thus avoiding the generating of numerous redundant class definitions for similar document subtypes. The resulting document management system DMS is thus capable of supporting the dynamic addition of new document types, and of uniformly processing queries spanning across multiple document types. In this paper, we also describe the implementation of our approach on the commercial DBMS system Illustra to demonstrate that the ease with which our approach can be realized on current OODB technology -- without requiring any special-purpose constructs. Our DMS system provides support for integrated querying of both structural as well as content-based predicates across arbitrarily complex document types.