The Semantic Web for Information Owners
IN THIS CHAPTER
• Precursors of the Semantic Web
Architecture of the Semantic Web
How Do Semantics Get into the Semantic Web?
What is the Semantic Web? Recall the fable of the blind men and the elephant, where the blind man grappling the elephant’s leg said, “An elephant is like a temple pillar! The blind man who lifted the elephant’s trunk said, “An elephant is a snake,” and the blind man who was brushed by the elephant’s flapping ear said, “An elephant has fronds like a palm tree!” In speaking of something of the scale and scope of the Semantic Web, we are like those blind men, and the Web is like that elephant, except that the Semantic Web is an elephant under construction.
Therefore, we must understand the Semantic Web under conditions of uncertainty. First, we’ll look at the ancestry of the Semantic Web, where you will see the two main lineages from which the Semantic Web inherits concepts: bibliography and knowledge representa-tion. Some of these ancestors are still alive today and meeting requirements, giving us confidence that the fundamentals of the Semantic Web are sound.
Then, we’ll look at the seven-layer architecture of the Semantic Web, as listed here:
Unicode and URIs
XML, XML Schema, and XML Namespaces
RDF, RDF Schema, and Topic Maps
When the Semantic Web is constructed, its central value proposition will be conversa-tions between people and machines. For example, take the following English sentence:
Buy me a hardcover copy of Jane’s book, in Chinese, if available for less than $39.95.
If this sentence is expressed using RDF statements and vocabularies (see Chapter 23, “RDF for Information Owners”), both machines and humans will be able to understand it. Furthermore, the machine (using the Ontology and Logic layers) will be able to draw inferences from the statement—and make further statements. For example, if the book is not available at the right price, the machine will be able to suggest an alternative book to the human as well as understand the human’s answer. Also, if the human doesn’t under-stand how the computer came to the conclusions it did, the computer’s logic can be exposed by the Proof layer. Finally, if asking the computer for proof all the time is just too time-consuming, one can ask trusted and more tractable humans (or agents or machines) what they feel. We collect all these question and answer interactions under the heading of conversation—realizing that this could end up making machines seem more human than we expect them to seem today. In fact, conversation is what puts the semantics in the Semantic Web.