Things in the Web need to be identified and this is usually done with an
Uniform Resource Locater (URL, i.e. web address) and the generalized form is the
Uniform Resource Identifier (URI). When we know how to identify things we want to represent facts about these things. This can be done with the
Resource Description Framework (RDF) which enables us to formulate sentences (really simple sentences) that represents facts about the world that we want to represent. It is not just about specialized individual facts, what need to be done is to come up with a model of the world. This can be done with RDF Schema language, which is a schema definition languages that enables us to model small worlds with explicit Semantics.
Of course, if we represent knowledge we need a way how to access this knowledge if we have a large knowledge base of this RDF facts. This in turn can be accomplished with the query language SPARQL, which is similar to SQL as it is used in Databases, but it is based on Semantics. Accessing large amount of data requires to think about how we can persistently store this data. This can be done with specialized RDF Databases like
Triple Stores. The Term Triple refers the structure of the RDF based Databases. These Databases contain facts and this facts are represented as Triples of subject, predicate and object in order to make a small sentence of a fact in the world.
If we now have a knowledge base somewhere in a Triple Store that represents a domain in RDF then it needs to be put somewhere on the Web. This can be done with technologies like RDF attributes (RDFa) and micro-data which is nothing else than semantic meta-data that can be put into the HTML Representation of a Web page.
However, the world is much more complicated than simple subjects, predicates and objects can represent. There are rules, restrictions and constrains which requires adequate Knowledge Representation.