Supporting Web content quality: Formalizing metadata concepts for the Web domain

Poor Web data quality can have many costs for the typical enterprise. This is true for Web-based data as wen as for traditional data sources. This applied research uses a series of prototyping and validation tasks to investigate the use of metadata, systems to support Web content quality. Two types of metadata are explored: content metadata that describes the purpose of a Web document, and dependency metadata, that describes the dependency relationship between content in different Web documents.; This research begins by exploring the use of metadata to provide specific information about the content of a Web page. A protocol suite for deploying and sharing content for a specific domain, Decision Support Systems (DSS) is proposed. A metadata (Open DSS) protocol that describe DSS such that appropriate DSS can be located and understood by decision-makers is developed and validated. The validation utilizes an experiment to assess end-users' understanding of specific DSS capabilities based on protocol compliant metadata descriptions. The validation methodology used would be appropriate for validating a variety of Web metadata, models.; Next, this dissertation addresses the problem of data consistency between related content available on the Web. In this phase of the research metadata concepts are expanded to develop a formal metadata model and language that allows the precise and detailed specification of complex Web object dependencies and their representations. It allows efficient strategies to enforce the consistency requirements among data contained on multiple Web sites. The dependency specification language, XDependency, is a XML-based language can be used to maintain Web-based data such that it meets businesses' consistency needs, reduces time-consuming and costly manual operations, and provides data of better quality to end users.; The concepts developed in this dissertation will facilitate the maintenance of Web-based content by allowing metadata about Web documents and the relationship between Web documents to be distributed with Web documents. Intelligent agents will be used to translate this metadata into appropriate action.