Wes Boudville
Hyperlinks and the Semantic Web

If you are like most, your introduction to hypertext was when you used a browser for the first time. You saw how intuitive it was that links were visible on a web page, and how easy it was to click on a link and be taken anyplace else on the web. The great acceptance of the web and its browsers was due in no small part to the ease of use, and the ease of writing HTML pages. Then after a few years, some deficiencies became apparent. HTML does not separate content from presentation. To do so, XML arose, with its user definable tags. But XML, by itself, does not take the concept of linking beyond a link being a one way connection between a single source and a single destination. Yet the concept of hyperlinks is much broader and older than the web's instantiation, having been worked on in the 1960s by Nelson and Englebart. This book expounds on how hyperlinks generalise the links in HTML. You can have multiple sources and multiple destinations. The links can be bidirectional. Given a destination, you can find the documents with sources that link to it. Currently, with web pages, you have to use a search engine to see who links to your pages. And no search engine reaches over 50% of the web. The book describes the concept of a linkbase: a database of links. It can be stored separately from the underlying documents that it references. This lets you annotate a document without changing it! Imagine the possibilities. The document could be on a CDROM, or on someone else's website. The implementation of these concepts is through XLink, XPath and XPointer. The authors use Backus-Naur Formalism to describe the grammar and they illustrate it with examples. My only quibble is that perhaps they could have written problem sets, for newcomers to cut their teeth on. If you want to see a possible future direction of the Semantic Web, have a gander at this book.