Enhancing semistructured data mediators with document type definitions

Mediation is an important application of XML. The MIX mediator uses Document Type Definitions (DTDs) to assist the user in query formulation and query processors in running queries more efficiently. We provide an algorithm for inferring the view DTD from the view definition and the source DTDs. We develop a metric of the quality of the inference algorithm´s view DTD by formalizing the notions of soundness and tightness. Intuitively, tightness is similar to precision, i.e., it deteriorates when “many” objects described by the view DTD can never appear as content of the view. In addition we show that DTDs have some inherent deficiencies that prevent the development of tight DTDs. We propose “DTDs with specialization” as a way to resolve this problem