In the Spring 2007 Web-Based Publishing course we implemented a first approach towards formalizing linking structures on the Web as part of a path towards Declarative Web 2.0. The idea of linkbases is to formalize a concept for annotating or connecting Web resources, and to treat these annotations and links as self-contained entities. A linkbase then simply is a database which stores these entities and can be used to enrich the Web browsing experience by providing more context when browsing the Web.

For example, when visiting a product site, there may be reviews about that product available on the Web, but the product site will not have links to them. In a linkbase-enabled Web, the browsers queries a linkbase when accessing a Web site about links connecting this resource. So if the linkbase contained a link between the product site and the review site, the browser can present this additional context to the user.

While the first version of the linkbase infrastructure used some ad-hoc XML interaction between clients and linkbases, a more standards-based approach should be based on Atom and the Atom publishing protocol for the interaction with a linkbase. The next version of the linkbase prototype will therefore be Atom-based and as a result will be easier to use and integrate into other Web-based architectures.