Integrating Web and database information for collaboration through explicit metadata

Loosely structured Web-based collections and more highly structured database resources have value to the extent that they convey meaning, understanding, and usable information. Metadata, or 'data about data' is the means of conveying this meaning, and thus it is an essential ingredient for successful collaboration involving diverse and distributed information. We characterize three dimensions of metadata and then develop a unified metadata representation that spans heterogeneous database systems. Our ubiquitous representation also incorporates structured documents and Web-based information. This is in contrast with the current approach of having a different metadata representation for each data model, such as relational, object-oriented, etc. Our extensible representation also supports knowledge-based representation of semantics. We believe this approach will substantially enhance collaboration over diverse information resources.