| Recently, XML has emerged as a major standard for the exchange and sharing of data over the Internet. Although XML documents are supposed to be self-describing and thus do not require any schema, schemas are useful to more efficiently process, query, and manage XML documents. Most of the proposals for schema specification, such as Document Type Definitions (DTD) and XML Schema, are grammar-based schema formalisms. Purely pattern-based schema formalisms, however, have not been investigated as an alternative or add-on for modeling schema constraints.; In this thesis, we introduce the concept of tree pattern constraints (XTPC) for XML documents as a pattern-based schema formalism founded on XPath. With tree patterns as the underlying formalism for addressing structural components in XML documents, XTPC provide an effective means to specify conditions on patterns that XML documents have to satisfy. Conditions include implication, absence, and co-occurrence of patterns. XTPC can be used as stand-alone schema formalisms or in conjunction with a grammar-based schema such as a DTD or XML Schema.; In order to get a better understanding of the properties and usage of XTPC, we study the formal reasoning about XTPC, in particular, the bounded implication and consistency problems. We also discuss different approaches to validating XML documents with respect to an XTPC specification. Furthermore, as an application of XTPC, we introduce an XPath optimization framework that applies the knowledge of XTPC. |
