Advantage of Semantic Data
The World Wide Web Consortium (W3C) has coined the term Semantic Web to describe the methods and technologies proposed to allow machines to understand the meaning (semantics) of information on the Web.
Of course, this means that the Web must contain semantic data to provide the advantages of a semantic search. Today Semantic Web technologies have found use/adoption among core specialized communities and organizations to a greater degree than the Web. It appears that practical constraints for adoption are less of an obstacle when domain and scope is more limited. However, particularly as use of a common ontology becomes more feasible, semantic databases and the Semantic Web will continue to grow.
Semantic data is assembled in a graph database that is unlike the more common relational and hierarchical databases that have some elements of data (nodes or tables) that are more important than others. A graph DB uses arbitrary object relations with no intrinsic importance.
- A Relational Model uses SQL data and query syntax and the Primary Key Value is searched with SQL.
- A Hierarchical Model uses XML data and a Unique Attribute Key search with XPath.
- A Graph Model uses RDF/XML data and a RDFS/OWL syntax semantic search with SPARQL.
In the Semantic Model:
- RDF (Resource Description Framework) is a simple language for expressing a RDF statement.
- RDF/XML is syntax to express a RDF graph as a XML document.
- RDFS (RDF Schema) extends the vocabulary of simple RDF.
- OWL further extends the RDFS vocabulary.(RDFS and OWL are namespaces on W3C)
- SPARQL (Search Protocol and RDF Query Language or “sparkle”) is a query language for semantic data similar to SQL.
An RDF or triple is a statement with three constituent parts: a subject, a predicate, and an object, such as the sky (subject) is (predicate or “link”) blue (object or property) In RDF/XML each part is defined by a URI (Universal Link Identifier). Where this becomes most confusing for those starting out with RDF is the concept that a subject in one RDF/XML can be an object of a property in another RDF/XML statement. It takes learning and experience to understand and use this concept.
Without semantic data, the owners of separate SQL sites would have to decide on a common data format to share information that they could both understand. Unfortunately, this requires human intervention.
In the semantic data model, the same base ontology (such as DCMI or FOAF) for expressing the meaning behind the data is used by different sites that makes it available for SPARQL searching.
OWL (Web Ontology Language) classifies things in terms of semantics, or meaning, where a class is a group of “individuals” that share common characteristics. A machine reader “knows” if an individual is a member of a class and its object properties or data properties related to that OWL class.
SPARQL queries RDF datasets, rather like SQL queries a relational database, except SELECT identifies the triples values selected on the RDF graph that will be returned in the results and WHERE identifies the triple patterns on the RDF graph to be matched against the RDF dataset.