Programming the Semantic Web

Chapter 2. Expressing Meaning

In the previous chapter we showed you a simple yet flexible data structure for describing restaurants, bars, and music venues. In this chapter we will develop some code to efficiently handle these types of data structures. But before we start working on the code, let’s see if we can make our data structure a bit more robust.

In its current form, our “fully parameterized venue” table allows us to represent arbitrary facts about food and music venues. But why limit the table to describing just these kinds of items? There is nothing specific about the form of the table that restricts it to food and music venues, and we should be able to represent facts about other entities in this same three-column format.

In fact, this three-column format is known as a triple, and it forms the fundamental building block of semantic representations. Each triple is composed of a subject, a predicate, and an object. You can think of triples as representing simple linguistic statements, with each element corresponding to a piece of grammar that would be used to diagram a short sentence (see Figure 2-1).

Figure 2-1. Sentence diagram showing a subject-predicate-object relationship

Generally, the subject in a triple corresponds to an entity—a “thing” for which we have a conceptual class. People, places, and other concrete objects are entities, as are less concrete things like periods ...