For everybody who is confused about the various RDF serializations/syntaxes

A nice Venn diagram can be found in Tim BL's spec of Notation3 at http://www.w3.org/DesignIssues/Notation3 (Appendix).

Here is also a short overview:

• RDF/XML was the original serialization developed by the RDF Core Working Group (at the time when XML took off). A RELAX-NG Schema was defined for validation.
  MIME Type: application/rdf+xml
  Character encoding can be set by the XML encoding attribute
  Usual file extension: .rdf

• N3 is an alternative compact textual syntax with various add-ons introduced by Tim BL already in 1998. It has a greater expressivity of standard RDF/XML. A subset with equal expressivity of RDF/XML was defined later by Dave Backett as Turtle (see below).
  MIME Type: text/rdf+n3
  Character encoding is always UTF-8
  Usual file extension: .n3
• N Triples is (only) used for the RDF Test Cases (it's unambiguous which is required for correct matching of testcases)
  MIME Type: text/plain
  Character encoding is limitted to 7-bit US-ASCII
  Usual file extension: .nt
• Turtle was introducd by Dave Beckett. By contrast to N Triples it supports abbreviations, @prefix, and grouping of multiple S-P pairs and objects which is also supported by N3, but it does not add features which cannot be expressed in RDF/XML.
  MIME Type: application/turtle and application/x-turtle
  Character encoding is always UTF-8
  Usual file extension: .ttl

The SPARQL graph patterns are similar to Turtle with additional variable tokens and literals allowed for subjects.