|
|||||||
|
Metadata PersistenceAnnotea. In Annotea, annotations are external to the documents annotated, and stored either locally or in annotation servers. Client applications are expected to store privately authored RDF content locally as well as exchange shared RDF content with annotation servers in the standard HTTP protocol, i.e. using POST, GET, PUT, and DELETE methods as mentioned previously.
Annotation servers serve as shared RDF databases with simple HTTP-based access control. Annotea does not suggest centralized metadata management due to scalability [72] and privacy problems. Instead, developers are expected to setup their own annotation servers forming a wide network of annotation servers that authors can choose freely depending on the specialties or convenience of the location of servers. Sesame. In Sesame, the Repository Abstraction Layer (RAL) hides the vendor specific database implementations of queries by providing stackable layers of abstraction, each of which helps the system to cache query results and hide implementation details through application interfaces. This structure permits developers to base Sesame on different database management systems such as DBMSs, existing RDF stores, plain RDF files, and other distributed RDF repository systems. RDFSuite. For metadata persistence, RDFSuite is deployed on top an object relational DBMS. Due to its object-relational support, the RDFSuite sample implementation uses PostgreSQL [82]. As the RDF validator parses and validates RDF description and schemas, it also fills in the predefined tables such as Class and Property, and creates new database tables and relationships for user defined schemas. Once the tables are created, the RDF loader populates newly created tables with RDF descriptions. A drawback of this approach is that DBMS may be overflowed with excessive number of tables because users can define infinite number of different RDF metadata. SAM. SAM and CMCS use Slide for resource and metadata persistence. Slide uses a relational database for metadata storage, and the local file system to store resources (i.e. images, Web pages). It indexes metadata in database tables using the file path portions (the file path after the hostnames) of URLs as database keys, and stores resources in a directory, serving as the content repository, in the file system where the file path portions are used as relative paths. Metadata can be in XML, which allows SAM, or CMCS, to store structured content to represent complex data formats. IKM. IKM requires a private DBMS system for the protection of sensitive content. User interface software deployed behind the Web interface accesses the local content private and channel it to users’ media players. A proprietary data model that represents the Rights Markup Extensions is used as DBMS data model. |
||||||
|