Major Description
currently it's only possible to get SPARQL results in application/rdf+xml (with content-type header: text/xml) for graph query results and application/sparql-results+json (with content-type header: application/json) for tuple query results. this should be improved to expose all available official mimetypes with proper content-negotiation (accept headers). it should be possible to override accept headers by URL params, so that all mimetypes are also available within a browser for testing. the implementation should make use of sesame's (later rdf4j's) writer registries (rdfwriter + tuplequeryresultwriter) so that additional writers (mimetypes) are available without touching the endpoint's code: for example: there's a json-ld writer for sesame (rdf4j) which just needs to be dropped into the server's lib folder to be available..