Sir Timothy John Berners-Lee, co-inventor of the World Wide Web with Robert Cailliau, had a dream: make the Web itself smarter. The idea, full of great promise in the late nineties, was that the Web could have its own taxonomy, a semantic system that could be understood by software agents for enhanced search capabilities.
It didn’t quite work out as expected. The top-down approach suggested by Berners-Lee was dislodged by a perfect storm: Google revolutionized search, while increases in bandwidth, storage, and processing, rather than making the Web more unwieldy, created Web 2.0. Now social networking, wikis, and peer-to-peer networks, combined with increasingly sophisticated search algorithms, have taken the focus off of the Semantic Web.
Bruce Spencer, research officer at the National Research Council (NRC) in Fredericton, N.B., says the idea of the Semantic Web was borne of the view that an explosion in static Web pages would create a host of problems.
“Now we have things like AJAX (e.g. Asynchronous JavaScript and XML), where a Web-site will upload little pieces of a page, so that it appears to be a high-functioning application. But 10 years ago pages were flat and unresponsive, and the idea was to make the Web more available and useful to people in the day-to-day world.”
It’s tempting to argue, then, that Web 2.0 has fulfilled some of the promise of the Semantic Web. One thing is certain: the work surrounding the Semantic Web itself has become more modest in terms of its overall ambition.
“Back in ’94 there were just a few browsers around, then people found that the Web was easy, and that the cost of getting in was really low,” says Spencer. “This fed Web 2.0. Now a lot of students understand Java, but also simpler scripting programming languages. They can go to a community college and build some pretty interesting applications.”
This is something that wasn’t predicted — that so much human intelligence could be applied to programming. Description logic has been around for decades, providing a formal way for representing desired information, but it was not a widely held skill. However, although it’s easier to get into the game now, that doesn’t make the challenges of the Semantic Web any less.
“With description logic you can infer hierarchies, and one concept leads to another,” says Spencer. “But you need to have taken a few AI courses to know how description logic works, then you need XML and OWL (e.g. Web ontology language). Add to that some background in DXML (e.g. Dynamic eXtensible Markup Language), and these people become harder to find.”
Alejandro (Alex) L