Research On Design Methods And Normalization Of XML

XML application is based on well designed XML Schema or DTD, so the design and normalization of XML are essential. At present, there is no modeling tool designed specially for XML. Making use of UML which is usually modeling for OOP or E-R model which is usually modeling for Relational Database to design XML, UML or ER model is converted to XML and sometimes the results can not be expressed fully or even changed to the semantics, and the difference between conversion methods will produce different XML, etc. The research on XML software reuse design focus only on one theoretical aspect, and practicality is weak. Functional dependency is the main reasons for that XML tree structure exists of data redundancy, which is similar to Relational Database. However, the research on XML normalization, such as functional dependency and XML Norm Form, is sparse and scattered. Therefore, this paper researches on both XML design and normalization, and the main work is as follows:(1) The paper researches on the modeling mechanism of UML and XML, and gives the feasibility analysis and the methods about UML modeling for XML; According to UML and XML semi-structured tree characteristics, the writing method of XML Schema which is based on function model sequence and the improving method are proposed.(2) E-R model is a very popular method modeling for relational database. The paper researches on the method of E-R model modeling for XML. The design method of DTD is proposed according to E-R model; there are difficulties in converting E-R model to DTD because of the difference between E-R model and XML structure, so the paper proposes ERA tree and an algorithm which changes E-R model into an ERA tree. Two connection traps come out easily according to E-R modeling, the two methods to solve the problem are proposed.(3) Software reuse is a goal pursued by programmers. The paper summarizes the four methods, which is data reuse, glossary reuse, XSLT standard library reuse and module reuse, and three encapsulation strategy in XML module reuse; Reuse extending is implemented based on parameter entity in DTD, which is put forward; It also gives reuse extending models, which is used to design reuse extending XML Schema.(4) XML normalization mainly resolve the problems result from eliminating semantic restriction and expression ability, such as data redundancy, inconsistencies, operation exceptions, and etc. The paper sums up some present results, especially functional dependency, XML paradigm and XML criterion rules, and gives method to eliminate data redundancy and normalization rules which is to solve the data redundancy problem caused by functional dependency and transmission dependency.(5) This paper gives several suggestions on recommendations and requirements as reference for the development of XML systems. Using the way given in the paper to successfully realize questions set in network general item bank system projects, the content model view automatically generated from XMLSpy2007 and the initial design based on the demand structure of the papers are very much the same, that is, using the methods to achieve the design requirements. The methods could be used to design other XML systems.