Web 3.0 is dead on arrival. In Sunday's New York Times, respected technology journalist John Markoff detailed the coming of Web 3.0 - the movement to imbue digital data with meaning so that it can be better understood by computers - and the blogosphere shot the idea down in cold prose."There's no story here," states blogger Robert Scoble, pointing to Valleywag's similar dismissal. "Web 3.0 does not validate," says blogger Nick Bradbury.

Publisher and Web 2.0 Summit co-founder Tim O'Reilly says Markoff's description of Web 3.0 really describes Web 2.0, although he endorses the idea that "building systems that combine human and machine intelligence is a huge part of the oncoming future."

That's a safe bet - the notion that people and machines will work better together as time goes on isn't exactly going out on a limb. But it's also a far rosier view of machine intelligence than is warranted by reality.

Web 3.0, also known as the semantic Web among those with a strong stomach for jargon, has far more to do with search technology startups struggling to emerge from Google's shadow and get funded than with anything else. These companies generally begin with the premise that search doesn't work and proceed to argue that some new technology will answer search queries with something better than a 20-page list of search results.

There's some merit to that argument and if you want a vision of the future of Google, look at search engines like the atrociously named Clusty.com, which subdivides search results by conceptual category. There are now, and will be, better ways to refine search results and they'll see more widespread adoption soon enough.

But machine intelligence is no match for natural ignorance, determined deception, and inherent ambiguity.

People are the main reason searches fail. They simply don't know how to craft effective queries and they're not eager to learn. Machines can help by being trained to guess what users mean, but users themselves don't necessarily know what they're looking for.

Markoff writes that "the Holy Grail for developers of the semantic Web is to build a system that can give a reasonable and complete response to a simple question like: 'I'm looking for a warm place to vacation and I have a budget of $3,000. Oh, and I have an 11-year-old child.' Under today's system, such a query can lead to hours of sifting - through lists of flights, hotel, car rentals - and the options are often at odds with one another. Under Web 3.0, the same search would ideally call up a complete vacation package that was planned as meticulously as if it had been assembled by a human travel agent."

That question is simply too vague for a machine to do anything with. A person would have to provide a lot more details to get a satisfactory answer from a computer. Moreover, there's no right answer. There are dozens, if not hundreds of vacation spots that would fit such loose criteria. The proper way to answer this question is by asking someone you trust, someone who knows your likes and dislikes, for a recommendation.

And if the person asking the question can frame it in specifics, why bother with parsing written sentences via computer when a series of menu selections from a structured database would do just as well? If houses can be purchased by selecting a series of structured fields like price, location, and number of bedrooms, travel packages can be dealt with the same way. No semantic Web is needed in many of the imagined use cases.

People are also willing and able to deceive machines, not to mention other people. The idea that the semantic Web will somehow see through the Web of lies promulgated by spammers, cyber crooks, and partisans of various worldviews is just naïve. Certainly machines can clarify the signal and filter the nose, but they have limits.

Garbage in still tends to result in garbage out, and the semantic Web isn't going to keep the garbage out of the system any more than MySpace is going to be able to keep out pedophiles.

Finally, while the semantic Web may prove useful in identifying facts and helping to provide some structure to the unstructured Web, machines still fall short of people when it comes to evaluating information.

A case in point: Among the many off-target recommendations made to me by Amazon.com, perhaps none is more of a miss than Kierkegaard's Relations to Hegel Reconsidered, by Jon Stewart. This philosophy text was recommended to me because I added The Daily Show with Jon Stewart Presents America to my Wish List long ago. Amazon's software just isn't smart enough to recognize that the philosopher Jon Stewart is not the same person as the Jon Stewart who hosts "The Daily Show."

So welcome Web 3.0. May you rest in peace.