Functional dependencies for XML : axiomatisation and normal form in the presence of frequencies and identifiers : a thesis presented in partial fulfilment of the requirements for the degree of Master of Sciences in Information Sciences at Massey University, Palmerston North, New Zealand

Abstract

XML has gained popularity as a markup language for publishing and exchanging data on the web. Nowadays, there are also ongoing interests in using XML for representing and actually storing data. In particular, much effort has been directed towards turning XML into a real data model by improving the semantics that can be expressed about XML documents. Various works have addressed how to define different classes of integrity constraints and the development of a normalisation theory for XML. One area which received little to no attention from the research community up to five years ago is the study of functional dependencies in the context of XML [37]. Since then, there has been increasingly more research investigating functional dependencies in XML. Nevertheless, a comprehensive dependency theory and normalisation theory for XML have yet to emerge. Functional dependencies are an integral part of database theory in the relational data model (RDM). In particular, functional dependencies have been vital in the investigation of how to design "good" relational database schemas which avoid or minimise problems relating to data redundancy and data inconsistency. Since the same problems can be shown to exist in poorly designed XML schemas 1 , there is a need to investigate how these problems can be eliminated in the context of XML. We believe that the study of an analogy to relational functional dependencies in the context of XML is equally significant towards designing "good" XML schemas. [FROM INTRODUCTION]