Maintenance of Human and Machine Metadata over the Web Content

Abstract

Abstract. Semantics over the Web content is crucial for web information systems, e.g. for effective information exploration, navigation or search. However, current coverage of the Web by semantics is insufficient. Web information systems mostly create their own content based metadata (e.g., identified keywords) and user collaboration metadata (e.g., implicit user feedbacks) in a form of information tags -structured information with semantic relations to the tagged content. By information tags web information systems build a lightweight semantics over the Web content, in which they can store knowledge and information about the content and interconnections between information artifacts of the content. Crucial problem of information tags lies in dynamicity of the Web whose content is continually modified. This together with influence of time can lead to invalidation of information tags which are closely related to tagged content. We address this issue via maintenance approach based on automatically and semi-automatically generated rules that respect changes on the Web and time aspect. The maintenance utilizes a rule-based engine which watches changes in the tagged content, identifies dependencies among maintenance rules and builds optimal strategy of rules application. We evaluate proposed maintenance approach in two domains -programing repositories and digital libraries, which use shared information tags repository.

Managing content, metadata and user-created annotations in web-based applications

“…We employ the annotation model that allows binding annotation to a particular document part. It has roots in a use case where users create annotations in client web-based systems by selecting a text part of the document and adding annotation content and properties by filling the showed form [6].…”