XML Documents Normalization Using GN-DTD

Abstract

Designing a well-structured XML document is important for the sake of readability, maintainability and more importantly to avoid both data redundancies and update anomalies. This paper proposes to improve and simplify XML structural design using a normalization process. To achieve this, Graphical Notation for Document Type Definition (GN-DTD) is used to describe the structure of XML document at the schema level. Multiple levels of normal forms for GN-DTD are proposed and the corresponding normalization rules to transform from poorly designed into well-designed XML documents. A case study is presented to show the application of these normal forms and normalization algorithm.

Formal Framework of XML Document Schema Design

“…This will save their time and effort as the process of database schema design is difficult (Arenas & Libkin, 2004) and take so much time. Under this assumption, we argue that to produce a "good" schema of an XML document for any application, we should first produce a graph conceptual model Document Type Definition (G-DTD) (Zainol & Wang, 2010) at schema level and then apply a normalization rule (Zainol & Wang, 2011a) to transform G-DTD into a normal form G-DTD' and finally convert the conceptual model G-DTD' back to the DTD schema.…”

“…The example of G-DTD is shown in Figure 4. Whereas, the second layer of XML-DM is the G-DTD normalizer layer which assists the user to design redundancy-free XML documents by normalizing the G-DTD on the basis of a set of normal forms (Zainol & Wang, 2011a). It will serve as a visual aid for the normalization process which is always easy to understand and interpret rather than a theoretical approach.…”

“…In this operation, a user is allowed to input a set of FDs. As a start the user is required to input a set of FDs and then it will be classified and grouped according to the FD's definition (Zainol & Wang, 2011a). The model will classify the set of FDs as a GFD, PFD and TFD on the basis of attribute nodes and simple element nodes in LHS and RHS of FDs.…”

“…The transformation from 1XNF to 2XNF is supported by the normalization algorithm and rules (Zainol & Wang, 2011a). This normalization algorithm is embedded as a backend process of the operations to help the user to achieve the best results.…”