The web is about publishing and accessing information. The semantic web extends this with machine-processable RDF data. This talk will describe publishing and accessing RDF using the infrastructure of the web. This is based on the exchange of RDF graphs, providing a simple building block for both publishing and collaborating applications.

We will show that current HTTP, is sufficient to provide the connection between RDF publisher and RDF application and that the query paradigm allows scaling from small collections of RDF statements to large aggregations and wrapped legacy applications.

An experimental system, Joseki, has been built on these principles; it provides a coarse-grained network access paradigm for extracting a subgraph from a larger, remote repository with the fine-grained API for manipulating RDF graphs provided by Jena and other toolkits.