The World Wide Web Gets Smart

In the Internet age, it’s hard to tell the difference between the next big thing and the next big flop.

For tech-savvy entrepreneurs hoping to ride the wave of the next version of the World Wide Web, called the Semantic Web, hype plays a key role in their financial futures. It’s not always easy to tell, really, how far off their promises are of an intelligent Web that will enable search engines that don’t just dumbly return word matches but instead actually understand your questions.

Getting your computer to understand you is at the heart of Barbara McGlamery’s work. She’s an ontologist — someone who defines how to categorize content for websites like Entertainment Weekly and Martha Stewart Living. McGlamery used the Semantic Web to help her answer one of society’s most pressing questions: who has Brad Pitt been in a relationship with?

It’s a surprisingly difficult question for a computer to understand. Getting one to comprehend such trivial — yet uniquely human — queries is at the heart of the Semantic Web. It is an attempt to help machines understand how humans make connections between cold data — and how they create meaning from those connections.

If you type “Who has Brad Pitt been in a relationship with?” into a search engine, you’ll get about two and a half million results — a Wikipedia article, tabloid headlines speculating about his past and current trysts, and thousands of photographs of him with various lovers. The data is all there, but human intervention is needed to figure out how it all fits together. A human needs to determine the timeline of Pitt’s relationships and to rank their emotional importance.

There are those who believe that computers will be able to think like humans someday. “It’s coming to the level that knowledge is being generated by machines,” claims Joseph Pally, founder of a software company called Z Cubes. “We are close to letting the machines make decisions. It’s happening already — they’re trading stocks!” he exclaims.

But many technology experts believe the Semantic Web is overhyped. “There’s even hype about the hype,” says Erik Manley, manager of digital media at General Electric.

Somewhere between the grandiose philosophizing about intelligent machines, build up about  “Web 20.0,” and the keystrokes of carpal-tunnel-stricken programmers, the Semantic Web is gaining legs.

The tools of the Semantic Web are straightforward. In order to get a computer to create meaning from data you need metadata — or in other words, you need data about data. Making the World Wide Web smart requires extra computer code. Every website, story, or photograph containing Brad Pitt needs a type of digital tag, the metadata, that lets a computer know that it’s about Brad Pitt.

The way web programmers communicate this metadata is through a computer language called RDF, or “Resource Description Framework.” According to World Wide Web inventor Tim Berners-Lee, it’s one of the fundamental building blocks of the Semantic Web. News organizations such as the New York Times have already started embracing RDF, using it to help make smart links between different stories. This would allow a reader to rapidly navigate through decades of stories about a certain region of the world, like Afghanistan — or a certain celebrity, like Brad Pitt.

While having every imaginable piece of trivia on Brad Pitt at your fingertips and a breeze to navigate is a compelling aspiration, it’s unclear how long it might take for companies to embrace the Semantic Web. As Berners-Lee pointed out in a public lecture in February 2009, twenty years ago companies didn’t think they needed websites. These days, they don’t see why the need metadata on their websites.

“If you have a system that works, or kind of works, why change?” asks Marc Ehrig, a search product manager at Thomson Reuters and member of the New York City Semantic Web Meetup. “It’s a powerful tool, but it might not apply in every case,” he explains. Like other members of the group, Ehrig has faced challenges getting money to build Semantic Web infrastructure at his company.

No one knows better than Marco Neumann the difficulty of communicating the potential of semantic technologies. Since founding the New York Meetup group in 2005, he has recruited almost 1,200 New Yorkers to look at how the Semantic Web can improve business.

“When I moved to New York there was no conversation at all — or very little — about the Semantic Web,” Neumann says. While he has made a lot of progress educating people, the key challenge he faces is convincing companies to spend money on metadata.

Right now, it’s difficult to quantify the benefits of the Semantic Web because so few companies are using it. Most can’t justify the investment of upgrading their conventional Web system.

“You always aim for the bottom line and the last thing you want to do is change your information technology system,” observes Neumann.

But according to Pally, once enough companies invest in the Semantic Web, it’s only a matter of time before the Web achieves a level of knowledge called wisdom — and the dawn of an age of intelligent machines. “I used to not believe it, but it’s happening. I’ve stopped saying it can’t happen,” he admits.