What is Knowledge?

XML Topic Maps: Creating and Using Topic Maps for the Web
Edited by Jack Park
Addison-Wesley
2002
ISBN 0-201-74960-2
544 pp., US$44.99

A sweet spot between philosophy and computer science is where you’d find this book. With an outstanding array of authors and editors, XML Topic Maps attempts to answer a very difficult question: “What is knowledge?” The authors gently walk the reader through the book, offering in-depth understanding of the topic and showing that the answer lies in the journey, not the destination. They ask questions such as, “How can we model a shoe’s “shoe-ness” such that a computer can distinguish a shoe from a non-shoe?” and “How can we use this methodology to add semantics to Web pages so that they display relevant information when a user searches the Web?”

Their answer is topic maps, which are able to represent knowledge. They display relevant information such that the user views it without increasing cognitive loads, avoiding information overload. Have you ever “googled” a phrase and been astonished at the amount information retrieved? You probably read through numerous pages of irrelevant information before finding what you were looking for. That’s where the Semantic Web comes in. Semantic networks, a term that has so much history in artificial intelligence, now has its counterpart in the Semantic Web. In a sense, the entire book is about the Semantic Web, which annotates information with constructs that the computer can read, enabling the computer to make a decision about what information is relevant. Topic maps represent a neutral envelope that can be used to describe the metadata. 

The authors of XML Topic Maps take turns demonstrating the problem domain and available solutions. The topics are relevant not only to XML Topic Maps but also the Resource Description Framework and any other knowledge representation construct.  In fact, two chapters address XTM and RDF similarities and differences and how we can integrate these two constructs together to make a more powerful knowledge representation construct. Before we can intelligently discuss metadata representation and annotation, we must understand ontology engineering’s underlying theory. The book covers ontological engineering in detail, including its various problems. A main problem the authors tackle is correctly modeling knowledge so that it conserves the meaning and makes the model reusable. The following two sentences are ontologically the same:

•      “Sam is a parent of John”;
•      “Joe is a boss of Kevin.”

The authors show the reader how to model such constructs via XTM and, more importantly, how to reuse these constructs later. They also demonstrate numerous open source tools that you can use to represent knowledge and create XTM or RDF constructs. Screen snapshots on these tools assist the reader in learning how to use available tools to represent knowledge. These constructs can then be transformed via Extensible Style Sheet Language Transformation to HTML for Web page viewing. 

RDF has been getting a lot of attention and is the preferred way to represent knowledge. The authors don’t make the claim that we should abandon RDF for XTM. In fact, they demonstrate that XTM and RDF can work in harmony and should be combined for better results. 

Jack Park ties up any loose ends in the last chapter by returning to the question that he raised in the opening chapter about adding semantics to Web pages. He defines ontology as “the study of being” and “a kind of vocabulary” that describes things in our universe. Park writes, “To be in a Semantic Web, we must develop ways to ‘put meanings’ on the Web.” However, he adds that the Semantic Web must evolve so that it can represent and apply shared meanings.

We can’t really answer the question “what is knowledge?” but we can certainly find ways to study and describe knowledge. 

Art Sedighiis a solutions architect at Platform Computing and a freelance writer. Contact him at sediga@alum.rpi.edu.