Research On XML Conceptual Model And Its Transformation

XML(eXtensible Markup Language) has been widely used as a popular format to represent and exchange data. Many organizations store data in XML, and many database systems also support XML database. But a suitable conceptual model of XML is lacked.The research in this thesis is based on MDA principle. First, a new conceptual model of XML, named XUML, is introduced, which is a platform independent model. The XUML model is based on UML class diagram, and inherits the most semantics and features of UML class diagram. Matching the essential features of XML Schema: elements nesting/reference and types reuse, the XUML model introduces some new concepts such as business component, generic Aggregation, association in business component and so on.A business component contains a hierarchical XUML model. Generic aggregation represents hierarchical relationship, namely nesting of elements. Association in business component represents the reference and the hyperlink relationship between elements. Type reuse has two kinds: for complex type and for simple type. In order to reuse types at many times and many places, the XUML model can be divided into two parts: content part and type part, and the classes in the type part can be referred by the classes in the content part.The following is the main contribution in this thesis: transformation from XML conceptual model to XML logical model which is specified as XML Schema. In order to generate a better XML Schema document, Profiles are defined to extend XUML. The Profiles aim at the important features of XML Schema, and enable the semantics more explicit.This thesis defines some mapping rules from extended XUML, to XML Schema at some aspects such as class, attribute, relationship and constraint. Model transformation is based on these mapping rules. Before the transformation, marks are added to XUML, and then an extended XUML is produced. During the transformation, the content part is transformed to content document, and the type partis transformed to type document, then the two documents should be combined by "include" schema combination method. After the transformation, the generated documents should be adjusted according to application requirements.At last, a simple implementation method using PowerDesigner is introduced. For improving the model design and the model transformation, an implementation plan is proposed, which is based on Eclipse platform and XML Schema Infoset Model.