The Resource Description Framework (RDF) is a language for representing information about resources in the World Wide Web. It is a syntax independent data model that may be serialised in a variety of concrete syntaxes. RDF is the core data format used on the Semantic Web.
The Resource Description Framework (RDF) is a language for representing information about resources in the World Wide Web. It is a syntax independent data model that may be serialised in a variety of concrete syntaxes. RDF is the core data format used on the Semantic Web.
Data Model
The data model for RDF is Graph based. Every fact in the data is represented as a Triple (or Statement) composed of a Subject, Predicate and Object. The Subject and Object can be thought of as nodes in the Graph while the Predicate is an arc in the Graph. Graphs are composed of a set of Triples, note that duplicates have no meaning.
Subjects must be resources which can either be named (using a IRI) or anonymous (known as Blank Nodes). IRIs are considered to be globally scoped while Blank Nodes are scoped to the Graph in which they appear.
Predicates must always be named resources i.e. IRIs.
Objects may be resources - IRIs/Blank Nodes - or they may be Literal values. Literal values are of three types:
- Simple Literals - These are just strings e.g.
"simple literal" - Literals with Language Specifiers - These are strings with the language of strings specified using standard language tags e.g.
"literal with language tag"@en-gb - Literals with DataTypes - These are values with a declared data type e.g.
"123"^^xsd:integer. XML Schema DataTypes are commonly used and supported by related standards like SPARQL
Syntaxes
RDF is a syntax independent model but has a variety of officially and semi-officially standardised concrete syntaxes which are used to serialise it.
Official Standards (W3C)
The following are official standards created by the W3C
- RDF/XML is the official XML based serialisation of RDF. RDF/XML is quite commonly used but has some drawbacks in being difficult for a human to understand or write by hand.
- NTriples is the official plain text format for serialising RDF. It is a very simple serialisation that uses no syntactic sugar, it is very quick to parse and can be read/written by hand. It's disadvantage is that it is very verbose compared to other syntaxes.
- RDFa is the officially recommended mechanism for embedding RDF into existing (X)HTML web pages. It is easy to use once learned but is sometimes criticised for the fact that the DOM structures required to embed RDFa often don't fit with the DOM structures of pages forcing developers to embed additional hidden elements with the RDFa.
- Turtle is a readable yet compact plain text serialisation of RDF developed by Dave Beckett who was one of the members of the W3C working group that created the RDF specifications. It is quick and easy to parse and includes various syntactic sugar for compressing the data to make it less verbose. Turtle is a superset of NTriples and a subset of Notation 3 Turtle was introduces as an official RDF serialization in February 2014.
- JSON-LD - JSON-LD is another JSON serialisation of RDF. It emphasises compact syntax and the ability to add RDF into your existing JSON data. It becamse official along with Turtle in February 2014.
Unofficial Standards
While the following are not officially standardised they are reasonably well specified and are widely implemented in RDF/Semantic Web APIs:
- Notation 3 is a readable yet compact plain text serialisation of RDF developed by Tim Berners-Lee. It is a superset of Turtle and is capable of expressing data which is outside of the core RDF model as well as expressing all valid RDF.
- Talis RDF/JSON - One of several competing JSON serialisations of RDF that tries to be as compact as possible
