04/25/2002

By DOUG BEDELL / The Dallas Morning News

Producers of the Web's newest search engine, Teoma.com, are quick to shirk the mantle of "Google-killer."

"It's not really us saying we're the biggest challenge yet to Google; it's other people," says Paul Gardi, Teoma's vice president, with an aw-shucks drop in his voice.At the same time, Mr. Gardi will tell you that his product – relaunched this month after making a largely ignored beta appearance Sept. 11 – is "the most advanced search technology out there today." And given time, he predicts, its advantages over the competition will become apparent to millions of Net users.

Teoma (the Gaelic word for expert, pronounced tee-O-ma) exploits one facet of the Internet that other search technologies ignore, its architects contend: deep, specialized Web sites created by experts. There may be 900 million Web pages currently on the Internet, but only a fraction are collections created by the world's leading experts in their fields.

These highly developed repositories are often hidden or ignored in results returned by other search engines, the Teoma argument goes. That's because Google, for example, primarily ranks results by considering how many times other Web sites link to them. The more links, the more authoritative the source, so the logic goes.

Teoma founders believe a better search starts with a more limited set.

Teoma first searches out clusters of highly regarded expert pages containing the search terms. It then uses the number of links within that expert community to rank returns, a la Google. The result, Mr. Gardi says, is a group of more relevant results and less "junk."

Each Teoma search produces three lists. One features Teoma's top-ranked pages. Another shows links to expert communities by specific topic. The third is a collection of links from the highest-ranked experts.

"Instead of getting broader and broader, it focuses in on communities gathered around the subjects that you're really interested in exploring," Mr. Gardi says. "We're the first technology that converges, rather than diverges. We converge on the information rather than looking backwards to see what's related."

Teoma's crowing aside, many Google addicts have found its interface and performance less than compelling.

Although its spare, multicolored, logo-dominated page is similar to Google's original look, Teoma lacks equivalent features. There is no image search. No older, cached pages are served up. It doesn't scan newsgroups as Google Groups does. There is no Teoma Tool Bar for your browser. And it doesn't suggest spellings as Google does when you've misspelled a search term.

Besides that, says Danny Sullivan of SearchEngineWatch.com, its database leaves it short of a Google-level performance. Google has searched or cached more than 1.5 billion Web pages. Teoma has archived about 200,000, even after being combined in December with the database of AskJeeves (Ask.com, "Your Internet Butler").

"Hopefully, Ask Jeeves will provide funding to deliver on the promise of Teoma, giving us yet another quality search engine," Mr. Sullivan says. "If not, then Teoma might be yet another hot tool that is ultimately forgotten."

Mr. Sullivan says Teoma performs quickly now, but only because it has less traffic.

"It has a limited database at the moment, which is occasionally apparent and a little annoying," Mr. Sullivan says. "It also, because of its approach, has a tendency to send you to Internet lists of information rather than direct to the original source.

"However," he says, "Teoma definitely shows promise. Besides, it'll be nice to see Google have some competition."

Teoma contends the number of pages doesn't matter.

"Only about 500 million or so are actually useful," says Mr. Gardi. "When you get 30 or 40 million pages on the search term 'Macintosh,' the last page in that list isn't very useful."

Because of that philosophy, Teoma's searches may not be for everybody.

"It may be way too much for Joe Blow who is looking for news on a local soccer team," Mr. Gardi says. "But if you're researching a really meaty, broad subject, we're the only technology that can find these kinds of pages."

Teoma's relaunch this month has at least provided a spark for search debate.

Internet discussion lists and other interactive forums are filled with individual assessments of its merits and shortfalls.

Says Ben Silverstein of DotCom Scoop.com: "Teoma's product is weak, very weak. A search for 'dotcomscoop' brought back 168 results. The results included individual pages on my Web site that haven't been live since December."

But the Teoma staff says that sort of criticism doesn't bother them – yet. Teoma's Web-scouring robots have only analyzed about 400 million Web sites so far. About 200 million pages, Mr. Gardi says, have proved unusable.

"We're new, and we're going to keep growing and finding these new communities until we feel we have reached the very end of the Web," he says.