Once a means for transforming the source data to RDF has been selected, a project team must decide what RDF serializations they wish to use for publishing their data. It should be noted that there are a variety of tools available for converting RDF between various serializations, so there is little need to fear about marginalizing a data set by choosing the "wrong" RDF serialization. The NCSU project team looked toward other linked data producers for guidance on what RDF serializations to use, looking at linked data sets from DBpedia, OCLC, and the Library of Congress. Ultimately, it was decided to write XSLT files that would convert source data to RDF-XML (http://www.w3.org/TR/rdf-syntax-grammar/), N-triples (http://www.w3.org/TR/n-triples/),Notation3 (N3)/Terse RDF Triple Language (Turtle)(http://www.w3.org/TeamSubmission/n3/, http://www.w3.org/TR/turtle/), and JavaScript Object Notation for Linked Data (JSON-LD) (http://json-ld.org/). Though JSON-LD is not technically an RDF serialization, JSON is a popular format among programmers that shares some syntax similarities with N3/Turtle, so a JSON-LD XSLT file was created rather easily from the template of the N3/Turtle XSLT file.