One great way to start a fight in a crowded Silicon Valley cocktail party (and there are a lot of them these days) is to mention Web 3.0.

There is no easy consensus about how to define what is meant by Web 3.0, but it is generally seen as a reference to the semantic Web. While it is not that much more precise a phrase, the semantic Web refers to technology to make using the Internet better by understanding the meaning of what people are doing, not just the way pages link to each other.

Amid the new Silicon Valley gold rush under way, a lot of entrepreneurs seem to believe that to define something is to own it. And Web 3.0 seems like a great thing to own.

So companies are bubbling up all over the place that claim to be building part of the semantic Web. Some are building voice recognition systems to use while browsing the Internet on a cell phone. Some want to challenge Google head on with a better search engine.

The leading players include Danny Hillis, the founder of Metaweb, Barney Pell of PowerSet, and Nova Spivack, the co founder of Radar Networks.

At the Web 2.0 conference on Friday Radar Networks will show off Twine, a service that uses semantic Web technology to improve sharing information with friends and coworkers.

Mr. Spivack, who previously founded Earthweb, an early Internet development firm, is the grandson of management theorist Peter F. Drucker. It was funded in part by Paul Allen, the co founder of Microsoft who appears to be creating a small keiretsu of semantic Web start-ups.

Twine will be available in a limited test version on Oct. 29 and open to the public next spring. The idea is to create a web, not of your friends as on a social network, but instead of all of your information. Twine is intended to let you suck in email, bookmarks, RSS news feeds, websites, photos, videos, database and any other digital information. Then it tries to make sense of it.

The company is positioning itself like a Facebook for sharing information, rather than entertainment, with friends and associates.

In a demonstration I saw earlier this week Twine appeared to do a good job of what artificial intelligence researchers refer to as "entity extraction," that is categorizing things like people and places automatically. So you will be able to find your stuff by typing in a category–job applications, Cape Cod beaches, and so on.

The service will succeed to the degree that it accomplishes its goal of exposing the meaning of the information and automatically revealing relationships to enrich information by discovering patterns that users might not otherwise identify.

In the past such "knowledge management" services have been restricted to large corporations and to world of government intelligence organizations. Now the falling cost of computing and networking will make it available to everyday consumers and in theory support it with advertising.

The program will organize data in a popular format used defined by the World Wide Web Consortium known as Resource Description Format, or RDF, in principle making it as easy to export to new kinds of services as to import it.

To be sure, there is one easy way to separate this sort of heavy duty computer problem, from the Web 2.0 chat programs and such, Mr. Spivack explained in an interview at the company's headquarters which are located a South-of-Market neighborhood in San Francisco.

"If you're looking to make a quick buck you wouldn't do this," he said .

http://bits.blogs.nytimes.com/2007/10/19/what-i-meant-to-say-was-semantic-web/