* Adaption of final RDF API as decided on in Nepomuk.
* Ported RDF repository to my Soprano changes.
* New identifier system which actually allows the generic URI approach to live side-by-side with arbitrary identifiers. Each resource gets a unique and artificial URI which has no relation to for example the path of a file. The latter will be stored as an identifier. Both the URI and the identifier can be used to access the resource later. This way we can keep on using the nice syntax of KMetaData resources in combination with the Nepomuk way of handling resource URIs.
* Added more tests.

TODO: * not all methods of the RDF repository are implemented yet.
* the identifier system mentioned above (based on the nice kickoffUriOrId) produced some quite confusing code. I think it would be a good idea so somehow clean it up. How, I don't know yet. The biggest problem probably is to keep it fast, i.e. keeping the number of DBus calls small.