Research On XML Design Methods

With the rapid development of Internet, all kinds of heterogeneous data exchange are more and more frequent. XML(eXtensible Markup Language) is widely used in various fields as one of the primary standards for data presentation and exchange on the Internet, such as Electronic Commerce, Web Service, Digital Library and Enterprise Application Integration(EAI) and so on. A well designed XML schema is very important for XML applications. There are many researches on design methods about XML, but they are dispersive and not systematized. It is a significant topic of making these methods into a system. This thesis studies the corresponding design methods of XML logical schema from structural and semantic aspects.The XML Schema design methods are summarized and the quality specifications of designing XML Schema are proposed in the first part of this thesis. The standards include: legibility, reusablilty, modularization, extensibility, maintainability and flexibility. Furthermore, the methods of guarantee quality and their applications are mainly discussed.Design patterns are important design methods in Object-Oriented fields. According to these ideas, six XML design patterns in the second part of this thesis are summarized and described. A unified description is given to these XML design patterns, including flyweight, reuse, composite, extensibility, fan trap and chasm trap. These patterns are reusable. If we use them when we design XML Schema, they can enable XML schema more flexible, enhance the quality of XML document and improve the design efficiency.While providing strict syntactic constraints, the related XML specifications provide little semantic content, and lack the ability of constraint expression. It results in the redundant and inconsistent data. There are some researches on the XML normalization of XML schemas now, but most of them lack of systematization. We study the normalization theory of XML schema design, such as functional dependencies, key, normal form and normalization in the third part. A series of normalization theories are summarized for XML schema.In the relational database design, conceptual models are designed firstly, then normalized, and transformed into logical models at last. In contrast, the normalization theory of XML mostly begins to analyze logical schemas and instance documents of XML, then normalizes the schemas. At the same time the normalization theory is so difficult that it is hardly applied to practice. So in the fourth part we show the process of XML normalization in XUML conceptual model. Finally, an example is presented to explain the general application of XML design methods.The integrated design methods are brought forward in this thesis by the research on the design methods of XML schema. According to the XML design quality, we start from design XUML conceptual model, then normalize XUML. The normalized XUML model is transformed into XML schema. The XML design patterns can be used during this design process. These methods are helpful in optimizing XML documents and XML database. Thus better XML applications will be developed.