Just in case you missed it, the web now has version numbers. Nearly three years ago, amid continued
hand-wringing over the dot-com crash, a man named Dale Dougherty dreamed up something called Web 2.0,
and the idea soon took on a life of its own. In the beginning, it was little more than a rallying cry, a belief that
the Internet would rise again. But as Dougherty’s O’Reilly Media put together the first Web 2.0 Conference in
late 2005, the term seemed to trumpet a particular kind of online revolution, a World Wide Web of the people.
Web 2.0 came to describe almost any site, service, or technology that promoted sharing and collaboration
right down to the Net’s grass roots. That includes blogs and wikis, tags and RSS feeds, del.icio.us and Flickr,
MySpace and YouTube. Because the concept blankets so many disparate ideas, some have questioned how
meaningful—and how useful—it really is, but there’s little doubt it owns a spot in our collective consciousness.
Whether or not it makes sense, we now break the history of the Web into two distinct stages: Today we have
Web 2.0, and before that there was Web 1.0.
Which raises the question: What will Web 3.0 look like?
Yes, it’s too early to say for sure. In many ways, even Web 2.0 is a work in progress. But it goes without
saying that new Net technologies are always under development—inside universities, think tanks, and big
corporations, as much as Silicon Valley start-ups—and blogs are already abuzz with talk of the Web’s next
To many, Web 3.0 is something called the Semantic Web, a term coined by Tim Berners-Lee, the man who
invented the (first) World Wide Web. In essence, the Semantic Web is a place where machines can read Web
pages much as we humans read them, a place where search engines and software agents can better troll the
Net and find what we’re looking for. “It’s a set of standards that turns the Web into one big database,” says
Nova Spivack, CEO of Radar Networks, one of the leading voices of this new-age Internet.
But some are skeptical about whether the Semantic Web—or at least, Berners-Lee’s view of it—will actually
take hold. They point to other technologies capable of reinventing the online world as we know it, from 3D
virtual worlds to Web-connected bathroom mirrors. Web 3.0 could mean many things, and for Netheads,
every single one is a breathtaking proposition.
Tim, Lucy, and The Semantic Web
The Semantic Web isn’t a new idea. This notion of a Web where machines can better read, understand, and
process all that data floating through cyberspace—a concept many refer to as Web 3.0—first entered the
public consciousness in 2001, when a story appeared in Scientific American. Coauthored by Berners-Lee, the
article describes a world in which software “agents” perform Web-based tasks we often struggle to complete
on our own.
The article begins with an imaginary girl named Lucy, whose mother has just been told by her doctor that she
needs to see a specialist. “At the doctor’s office, Lucy instructed her Semantic Web agent through her
handheld Web browser,” we read. “The agent promptly retrieved information about Mom’s prescribed
treatment from the doctor’s agent, looked up several lists of providers, and checked for the ones in-plan for
Mom’s insurance within a 20-mile radius of her home and with a rating of excellent on trusted rating services.”
That’s quite a mouthful, but it only begins to describe Berners-Lee’s vision of a future Web. Lucy’s Semantic
Web agent can also check potential appointment times against her mother’s busy schedule, reschedule other
appointments if need be, and more—all on its own, without help from Lucy. And Lucy is just one example. A
Semantic Web agent could be programmed to do almost anything, from automatically booking your next
vacation to researching a term paper.
How will this actually work? In Berners-Lee’s view, it involves a reannotation of the Web, adding all sorts of
machine-readable metadata to the human-readable Web pages we use today (see “Questions of Semantics,”
opposite). Six years after the Scientific American article, official standards describing this metadata are in
place—including the Recourse Description Framework (RDF) and the Web Ontology Language (OWL)—and
they’re already trickling into real-world sites, services, and other tools. -Semantic Web metadata underpins
Yahoo!’s new food site. Spivack’s Radar Networks is building a kind of Semantic Web portal. A development
platform, Jena, is in the works at HP. And you’ll find Semantic Web structures in Oracle’s Spatial database
The problem is that a complete reannotation of the Web is a massive undertaking. “The Semantic Web is a
good-news, bad-news thing,” says R. David Lankes, an associate professor at Syracuse University’s School of
Information Studies. “You get the ability to do all these very complex queries, but it takes a tremendous
amount of time and metadata to make that happen.”
The Other Semantic Web
As a consequence, many researchers take a very different approach to the Semantic Web. Rather than
calling for an overhaul of Web formats, which would involve hundreds of thousands of independent sites,they’re building agents that can better understand Web pages as they exist today. They’re not making the pages easier to read, they’re making the software agents smarter.
One early example is the BlueOrganizer from AdaptiveBlue (www.adaptiveblue.com).
In certain situations, when you visit a Web page, this browser plug-in can understand
what the page is about, automatically retrieving related information from other sites
and services. If you visit a movie blog, for instance, and read about a particular film, it
immediately links to sites where you can buy or rent that film. “It’s what you might call a
top-down approach,” says Alex Iskold, the company’s CEO. “Web pages already
contain semantic data. We can understand them, so why shouldn’t computers? Why not build a technology
that can parse and process existing services and databases?”
Of course, that’s easier said than done.
Countless companies offer tools similar to BlueOrganizer—including Claria’s PersonalWeb—but these aren’t that different
from the old Amazon.com “recommendation engine,” which suggests new products based on
your surfing and buying habits. We’re a long way from agents that can think on their own. In the
near term, the Semantic Web may require the sort of metadata Berners-Lee proposes.
“Automated agents are worth striving for,” says Pattie Maes, an MIT Media Lab veteran who
founded the Lab’s Software Agents Group. “But it’s hard to say what’s better—tags built into Web
pages or tags that are, in a sense, inferred by machines.”
Semantics and Search
The Semantic Web, like Web 2.0, is a nebulous concept. “Considering that the very word
semantic is all about meaning, it’s ironic that the term Semantic Web is so ill defined,” says Radar Networks’ Spivack. Some, like Spivack, fall into the Berners-Lee camp. Others, like AdaptiveBlue’s Iskold, believe in the artificial-intelligence method. And then
there are the others: the semantic searchers.
Rather than providing automatic information retrieval, semantic search engines seek to improve on the
Google-like search model we’ve grown so accustomed to. The idea is to move beyond mere keyword
searches to a better understanding of natural-language queries. “Right now, search engines can’t tell the
difference between Paris Hilton and the Hilton in Paris,” says Jeff Bates, cofounder of Slashdot, one of the
driving forces behind Web 2.0. “There’s millions of dollars being spent trying to better optimize search, and
that’s a big part of what the Semantic Web will be.”
This kind of natural-language processing has been in development for years, but it, too, has found its way
onto the public Web. Several start-ups, including Powerset and TextDigger, are hard at work on semantic
search engines based on the open-source academic project WordNet. It should be noted, however, that
natural-language search could very well play a role in the Berners-Lee Semantic Web. His is merely a
framework to enable all sorts of apps, and semantic search might be one of them.
A Web Beyond Words
Though Web 3.0 is most often associated with the Semantic Web, the two are far from synonymous.
Countless other concepts are poised to play a role in our online future, and many go beyond semantics, using
space, images, and sound.
One possibility is the so-called 3D Web, a Web you can walk through. Many see this as an extension of the
“virtual worlds” popping up on today’s Internet. In the future, they say, the Web will be one big alternate
universe reminiscent of Second Life and There.com. But others scoff at this notion, claiming it’s just a
less-efficient version of today’s Internet. They see the 3D Web not as an alternate universe but as a
re-creation of our existing world. On the 3D Web, you could take a virtual stroll through an unfamiliar
neighborhood shopping for houses or visit famous sites you’ve never seen. Google Earth already offers an
experience not far removed from this. “Today, with a service like Google Earth, you can zoom in on Seattle
and see how tall the buildings are,” says Syracuse University’s Lankes. “It really isn’t that much of a leap to
actually put you, or your avatar, in Seattle and let you walk around.”
The trouble is, 3D only goes so far. It doesn’t enhance the very 2D world of words, pictures, and video. For
many, the more interesting idea is a mediacentric Web, offering not just language-based search but pure
media search. Today we depend on keywords even when searching for images, videos, and songs—a
woefully inadequate system. Companies like Ojos and Polar Rose are working to reinvent media search,
hinting at a world where we search for media with other media—not just keywords (see “Look Ma, No
Then there’s the Pervasive Web, a Web that’s everywhere. Today’s Web already extends beyond the desktop,
to cell phones and handhelds, but it might extend even further—into our everyday surroundings. At the MIT
Media Lab, Maes is toying with the idea of Web-connected bathroom mirrors. As you brush your teeth in the
morning, there’s the latest news. Meanwhile, with his blog, the End of Cyberspace, Alex Soojung-Kim Pang of
the Institute for the Future envisions the Web automating much of what goes on in the home. Your windows,
for instance, could automatically open when the weather changes. With help from mesh networks—wireless
networks consisting of tiny nodes that can route data to and from almost anywhere—the possibilities are
Tomorrow’s Web, Today
In some respects, Web 3.0 is nothing more than a parlor game. Ideas tossed out here and there. But at the
very least, these ideas have roots in current trends. Many companies, from HP and Yahoo! to Radar
Networks, are adopting official Semantic Web standards. Polar Rose and Ojos are improving image search.
Google and Microsoft are moving toward 3D. No one can predict what Web 3.0 will look like. But one thing’s
for sure: It’ll happen.
An Idiot’s Guide to Web 3.0
What will Web 3.0 look like? Who knows? But here are a few possibilities.
The Semantic Web
A Web where machines can read sites as easily as humans read them (almost). You ask your machine to
check your schedule against the schedules of all the dentists and doctors within a 10-mile radius—and it
The 3D Web
A Web you can walk through. Without leaving your desk, you can go house hunting across town or take a tour
of Europe. Or you can walk through a Second Life–style virtual world, surfing for data and interacting with
others in 3D.
The Media-Centric Web
A Web where you can find media using other media—not just keywords. You supply, say, a photo of your
favorite painting and your search engines turn up hundreds of similar paintings.
The Pervasive Web
A Web that’s everywhere. On your PC. On your cell phone. On your clothes and jewelry. Spread throughout
your home and office. Even your bedroom windows are online, checking the weather, so they know when to
open and close.
Questions of Semantics
Tim Berners-Lee isn’t the only man behind the Semantic Web. His 2001 Scientific American article, which
introduced the concept to the world, was actually written in collaboration with two other eminent -researchers,
Ora Lassila and Jim Hendler. Six years on, we tracked down Professor Hendler, now director of the Joint
Institute for Knowledge Discovery at the University of Maryland and still one of the driving -forces behind this
Q: Does the Semantic Web idea predate your now-famous Scientific American article—or was that the
A: That’s the first time the term was coined and printed in a fairly accessible place. Recently, we’ve been
looking for the absolute earliest use of the term Semantic Web, and it seems to go a bit further back, to a few
small things Tim had written. He and some colleagues were using it locally within MIT and the surrounding
community in the late nineties.
Q: The Semantic Web can be a difficult concept to grasp. How do you define it?
A: What the traditional Web does for the text documents in our lives, the Semantic Web does for all our data
and information. Today, on my Web page, I can build a pointer to another Web page. But I can’t link data
together in the way I can link pages together. I can’t point from a value in one database to some other value in
some other database. To use a simple example, if your driver’s license number is in one place and your
vehicle identification number is in another, there should be a way of linking those two things together. There
should be a way for machines to understand that those two things are related.
Q: Why is this so necessary?
A: Right now, it’s very difficult to browse data on the Web. I can use a search engine that gives me the results
of a query and draws them as a list, but I can’t click on one of those values and see what it really means and
what it’s really related to. Today’s social networking is trying to improve this, with things like tagging. But if you
typed “polish” and I typed “polish,” how do we know we’re talking about the same thing? You might be talking
about a language and I might be talking about something that goes on furniture. On the other hand, if those
two names are precisely identified, they don’t accidentally overlap and it’s easier to understand the data we’ve
published. So the technology of the Semantic Web is, in a sense, the technology of precise vocabularies.
Q: And this, in turn, would allow a machine to go out across the Web and find the things we’re looking
A: Yes. It’s very hard for this to happen with just language descriptions. Our idea is to have machine-readable
information shadowing the human-readable stuff. So if I have a page that says, “My name is Jim Hendler.
Here’s a picture of my daughter,” the machine realizes that I’m a person, that I have a first name and a last
name, that I’m the father of another person, and that she’s a female person. The level of information a
machine needs would vary from application to application, but just a little of this could go a long way—as long
as it can all be linked together. And the linking is the Web part of the Semantic Web. This is all about adding
meaning to the stuff we put on the Web—and then linking that meaning together.
Look, Ma, No Keywords!
Three new Web services reinvent the way we look for music and images.You won’t search for media with keywords in the future-—you’ll search for media with media. To find an image, you’ll supply another image. To find a song you’ll supply another song. Don’t believe it? Three new services—image-crunchers Like.com and Polar Rose, and music-matchmaker Pandora—have already taken the first steps toward this new breed of media search.
Today, when you search the Web for music and images, you’re merely searching for the words that surround
them. When you visit Google Image Search and type in “Steve Jobs,” you aren’t really looking for photos of
Apple’s CEO. You’re looking for filenames and captions that carry those keywords—”Steve” and
“Jobs”—hoping the right photos are somewhere nearby.
There’s a sizable difference between the two. On any given image search, Google turns up countless photos
completely unrelated to your query, even as it misses out on countless others that may be a perfect match. In
the end, you’re relying on Web publishers to annotate their images accurately, and that’s a hit-or-miss
The situation is much the same with MP3s, podcasts, and other sound files. When trolling Web-based music
services, you can run a search on “Elvis” or “Jailhouse Rock.” But what if you’re looking for music that sounds
like Elvis? Wouldn’t it be nice if you could use one song to find other similar songs?
Ojos and Polar Rose are tackling the image side of the problem. Last spring, Ojos unveiled a Web-based
photo–sharing tool called Riya, which automatically tags your pictures using face recognition. Rather than
manually adding “Mom” tags to all your photos of Mom, you can show Riya what she looks like, and it adds
the tags for you. The service is surprisingly accurate, gaining a huge following from the moment it hit the Web,
but Ojos quickly realized that the Riya face-rec engine—which also identifies objects and words—could be
used for Web-wide image search.
That’s a mammoth undertaking, but, with an alpha service called Like.com, the company is already offering a
simple prototype. Today, Like.com is little more than a shopping engine. You select a photo of a product that
best represents what you’re looking for, and the service shows all sorts of similar products. But it’s an
Meanwhile, Polar Rose (www.polarrose.com) recently introduced a browser plug-in that does face recognition
with any photo posted to any Web site. For the moment, it’s just a means of tagging images
automatically—much like Riya. But unlike Riya, it already works across the length and breadth of the Net.
The closest equivalent when it comes to audio is Pandora, from a group of “musicians and music-loving
technologists” called the Music Genome Project. Since its inception in 2000, the group has analyzed songs
from over 10,000 artists, carefully notating the music makeup of each track. Using this data and a list of your
favorite artists, Pandora can instantly construct a new collection of songs that suit your tastes. Again, this is
hardly a Web-wide search engine, and unlike the image services from Ojos and Polar Rose, it relies heavily
on up-front human input. But it’s a step in the right direction. True media search is closer than you think.
BY Cade Met ,