Knowledge-based data management for digital libraries

摘要

Digital libraries provide services for accessing, manipulating, and displaying collections. The services typically include ways to query the context associated with the collections (metadata attributes), as well as mechanisms to display the digital objects that comprise the collection. We discuss the characterization of knowledge associated with a collection, and discuss mechanisms for managing data, information, and knowledge. In a knowledge-based digital library, data objects are discovered by mapping from scientific domain concepts, to the attributes used within a data grid collection, to the digital objects residing in a storage system. These ideas will be explored in the context of large-scale scientific data collections, and illustrated based on infrastructure under development at the San Diego Supercomputer Center.