XML conceptual modeling using UML

Abstract

The eXtensible Markup Language (XML) is increasingly finding acceptance as a standard for storing and exchanging structured and semi-structured information. With its expressive power, XML enables a great variety of applications relying on such structures - notably product catalogs, digital libraries, and electronic data interchange (EDI). As the data schema, an XML Document Type Definition (DTD) is a means by which documents and objects can be structured. Currently, there is no suitable way to model DTDs conceptually. Our approach is to model DTDs and thus classes of documents on the basis of UML (Unified Modeling Language). We consider UML to be the connecting link bet- ween software engineering and document design, i.e., it is possible to design object-oriented software together with the necessary XML structures. For this reason, we describe how to transform the static part of UML, i.e. class diagrams, into XML DTDs. The major challenge for the transformation is to define a suitable mapping reflecting the semantics of a UML specification in a DTD correctly. Because of XML's specific properties, we slightly extend the UML language in a UML-compliant way. Our approach provides the stepping stone to bridge the gap between object-oriented software design and the development of XML data schemata.