Do we need the Semantic Web?

"These choices have no actual connection to the real application, yet they are baked into the system," he added. "Anyone who uses the data has to know what these decisions are."

Key goals for the Semantic Web architects include reuse of data and what backers call "recombinant effects."

They hope that by letting computers digest and exchange information about context and meaning — a word that raises the hackles of artificial intelligence critics — they will allow data to survive the systems where it originated and traverse different applications as easily as browsers traverse the Web's billions of pages today. As that data takes on a virtual life of its own, it could be exploited and combined in unexpected and unexpectedly profitable ways, the thinking goes.

"The really exciting thing isn't that you can merge your own data between applications — that's like links on your own Web site," Berners-Lee said. "The really exciting thing happens when others have their data in a mergeable format and make it available. When that public information becomes mergeable, we're in for the next, very pronounced stage of Web evolution."

Security worries
That brave new world of interchangeable data — "exposing data hiding in documents, servers and databases," in Miller's words — elicits both scepticism and alarm from critics of the emerging project.

One concern is that businesses with a Semantic Web presence may have a new headache in trying to prevent information from being unintentionally shared.

"We don't want to have this universal network of knowledge that makes everything accessible to all parties," said the Burton Group's O'Kelly. "Companies need to be circumspect about disclosure."

The W3C, acknowledging concerns about corporate and personal privacy, says it plans a Semantic Web rules system for information sharing. The consortium is calling for position papers by 18 March for its workshop on rule languages for interoperability, set for 27-28 April in Washington DC.

Even though crucial protocols are still in the idea phase, the W3C is insisting that the Web's next big evolutionary shift has already begun.

The W3C's Miller devoted much of his keynote address — titled "The Semantic Web is Here" — to existing examples of Semantic Web technologies being developed or rolled out by major companies.

Nokia, for example, maintains long-standing Semantic Web activity of its own and has made its Semantic Web toolkit, known as Wilbur, available on the SourceForge.net open source development site.

Miller hailed the way Nokia has used Semantic Web specifications, particularly RDF, in its Series 60 phones and in its developers' forum. In one of Miller's examples, RDF metadata, or data about data, lets phones communicate to Web sites about how much bandwidth they have. In another, RDF lets Nokia automatically serve pages individually tailored for developers of particular applications for certain phones.

Miller also cited other examples: HP's use of Semantic Web technologies in its work building an online education resource for the government of Singapore; the IBM Internet Technology Group's development of Semantic Web applications, especially those in the life sciences; Adobe's addition of RDF-based XMP (Extensible Metadata Platform) in its Creative Suite, which Adobe says sits on more than 700,000 computers; and Oracle's inclusion of the RDF Network Data Model in its Oracle Database 10.2, due out later in the year.

Also during his keynote, Miller laid out plans to spread the Semantic Web religion. He said he plans to ask the W3C membership to endorse a working group devoted to Semantic Web education and communication, and he also plans a Semantic Web symposium for CTOs and CIOs June 22-24 at a yet-to-be-determined West Coast location.

After years of being called artificial-intelligence throwbacks with their heads in the clouds, Semantic Web backers point to these real-world implementations with evident satisfaction.

"The Semantic Web is starting to take off now," Berners-Lee said. "It is not yet so developed that [implementers] keep bumping into people doing related things yet — we are not yet really seeing the benefit of application areas being connected together in unexpected ways. But in certain areas, the critical mass has been passed. At the recent Semantic Web and life sciences workshop... there was serious excitement about the opportunities in integrating across life science disciplines, like genomics, proteomics, clinical trial and epidemiological data and so on."