Capturing Real World Derived Information Relevancies In XML

XML data dependencies play important roles in eliminating XML redundancies and normalizing XML schemas. The main existing XML data dependency models can be classified into two categories: the tree tuple based models and the path based models. The main considerations in these models are the structural relevancies of XML nodes but not the information relevancies in XML data which derive from the real world. Therefore, on the one hand, they cannot capture the essential XML data dependencies faithfully and completely; on the other hand, they have expressive deficiencies and semantic deficiencies in expressing XML data dependencies. In this paper, we adopt a novel notion and propose an XML data dependency model (XDDM) based on entity segments. Our main objective is to make XML data dependencies in XDDM match the information relevancies derived from the real world. Firstly, we summarize these information relevancies into three categories and propose the completeness condition of XML data dependency model. And then, the XML attribute dependency and the XML entity dependency are defined to compose the XDDM. Finally, we analyze the capabilities of XDDM in fulfilling the completeness condition and eliminating expressive deficiencies and semantic deficiencies.