Despite the evolution of complex storage facilities for different application domains (e.g. multimedia-enabled databases, or RDF triple stores), the file system is still the major repository paradigm for unstructured content that is created, stored and managed by end-users. Many desktop file systems support the addition of metadata to files, but there exists no platform-independent agreement on how to retrieve and manage metadata, be it in a user's private storage or in shared data repositories. In this paper, we point out why the file system has significant drawbacks regarding efficient management of metadata, and how extensions to the well-known WebDAV protocol can be designed so that metadata can be retrieved, stored, managed and utilized for increased user benefit in searching and browsing situations. We describe the drawbacks of traditional file systems and strategies how to overcome their limitations. Finally, we express ideas regarding how the file system API for end user applications can be extended so that applications are enabled to annotate content with attention metadata, and describe how we plan to realize these ideas in the course of the SemDAV project.

Top