The Semantic Web - Social Semantics: The Search for Meaning on the Web - page 63

Fig. 3.4 The Semantic Web stack

Inspired by the Principle of Least Power, he envisaged that each language
would extend and build upon lower-level languages. On top of RDF, Berners-Lee
envisaged a whole stack of more expressive languages being constructed. Although
the vagarities of the standardization process have caused various changes in the
'Semantic Web stack' and numerous conflicting versions exist, the original and
most popular version of the Semantic Web stack is given in Fig. 3.4 (Gerber et al.
2008). The W3C has commenced standardization efforts in a number of these
areas, and research in almost all levels of the stack has begun. The majority of the
research has focused on extending the Semantic Web with “ontologies” based on
description logic like OWL. As should be suspected given their heritage in artificial
intelligence, most of the work in description logic applied to OWL has focused
on determining the most expressive possible language that preserves decidable
inference. OWL itself works well with the Open World Principle, since it only
makes an inference by adding inferred statements and classifications, and so remains
monotonic. While almost any possible triple is acceptable in RDF, OWL allows
users to design ontologies that can even add constraints, such as cardinality and data-
typing, that can make some RDF triples inconsistent with a given OWL ontology.
Another part of the Semantic Web, originally unforeseen, is the query language
SPARQL , a query language for RDF similar to the popular database query language
SQL (Prud'hommeaux and Seaborne 2008). Current work is focused on Rule
Interchange Format ) (RIF), a rule-language similar to Prolog for both serializing
normal rules and operating over RDF data (Boley and Kifer 2008). Other higher-
levels on the Semantic Web stack such as 'Unifying Logic' remain mysterious, if
poetic and evocative.