Semantic Web Technologies

Microformat

A microformat is a very focused set of attributes of a single concept. Microformats do not typically start as the result of a standards committee, but are adopted because they fit a need. A venerable example is vCard, which captures the kind of name and contact information found on business cards.

Ontology

An ontology is a formal definition of concepts and the relationships between them for some specific domain of discourse. It defines a consistent vocabulary for assertions and queries in that domain. An ontology can be as basic as a term hierarchy. A company might have an ontology for its internal operators that defines EmployeeStockOptionPlan as a type of IndividualBenefit, which is a type of Compensation. The ontology can also specify more than that. For example, there is a relation vested-for between StockOptions and Employee. StockOptions have an issueDate (a Date) and a strikePrice (a Currency, value > 0). The assertion StockOptions vested-for Employee is true when issueDate is at least six years ago. The definition of Employee ensures that the person who was granted options is still employed at this company, on active status, and that they have not become a contractor. This ontology provides a consistent terminology that the company intends to use, and operational meanings for each term.

OWL (Web Ontology Language)

Yes, it should be WOL, but the standards committee decided they liked OWL better. Ontologies expressed in OWL define concepts and the relations (predicates) between them, for use in RDF files. For example, one ontology might define the purchases relation to have a subject that is a Human or a Company, an object that is a Product, and is non-transitive and non-symmetric. The ontology is useful for validating statements, enabling the inference of new statements, and documenting what the creator of the RDF file meant by the predicates they used.

RDF (Resource Description Framework)

RDF is a foundation technology of the Semantic Web; it is the way assertions are encoded. A fact in RDF contains a subject, a (binary) predicate, and an object. A few examples:

 ESOP23 is-instance-of EmployeeStockOptionPlan  
 ESPO23 issue-date "April 15, 2004".  
 ESOP23 vested-for JohnDoe.

There are several notations for RDF. The one used by machines is XML based, and all references are URIs (unique references to Web resources). RDFa is a version that fits better in HTML pages, and there are several human-oriented notations.

SKOS (Simple Knowledge Organization System)

SKOS is a standardized RDF markup for organizing and indexing a set of terms or concepts. Terms are related to other terms in a hierarchical sense. Terms can also be related in a more semantic sense. For example: A SKOS file might define tree as a concept with a label of “Tree” in English and “Arbol” in Spanish. Tree is a narrower concept than Plant, and is related to Shrub. For data integration, the Tree concept might declare that it is a close match with the Spermatopsida concept in a specialist biology vocabulary.

SPARQL (SPARQL Protocol and RDF Query Language)

The query language for triples. It is modeled after SQL, but adapted to use with data that is represented as a graph of triples rather than tables, columns, and rows.

Triple

A common term for an RDF assertion is triple, since each has three parts—subject, predicate, and object.

Triple store

A collection of RDF facts (triples) in some storage container. The Semantic Web equivalent of a database.