Recently, XML has been applied as a format for data representation on different levels of information systems. As the amount of data represented in XML grows, a need for methods as well as tools for modeling XML data grows as well.

Recently, XML schema languages like XML Schema are applied to describe XML data on the logical level. However, modeling XML data directly with these languages is a time-consuming and error-prone task as we discuss in detail in [1]. XML schemes are usually interrelated with each other and they are also related to other components in the system (such as relational databases). However, these interrelations are difficult to capture and maintain on the XML schema level.

As we show in [1], designing XML data first on the conceptual level could help to solve the problems of XML schema languages. For this, existing conceptual models such as the Entity- Relationship model (ER) or the model of UML class diagrams (UML) can be applied. However, several extensions to these basic models are required to cover the special features of XML data. There are fairly recent surveys of this area, concretely [2] (ER-based approaches) and [3] (UMLbased approaches).