RealTime IT News

Yahoo Exec: 'People Don't Want to Search'

Demo conference
(L to R) Howard Bloom, Peter Norvig, Prabhkar Raghavan, Jon Udell. Click to enlarge.
Source: Brian Solis, for DEMO
SAN DIEGO -- Search has become so pervasive on the Web as part of our regular activities that it's tempting to assume it will be central to any future Web advances.

Prabhakar Raghavan, head of research at Yahoo, didn't get the memo.

"It's very limiting to think of the Web as merely a mechanism for retrieving data," Raghavan said in a panel discussion here Tuesday wrapping up the two-day DEMO conference. "The next step is divining the intent of what people are doing and fulfilling their tasks. I've always believed people intrinsically don't want to search, they come to work to do what the boss wants.... We have to get further on achievement than the notion of the retrieval engine as the ultimate target."

One example Raghavan gave of where search is limited is trying to plan a vacation. "You spend hours or days from start to end with repeated search engine queries. At the end there's a glaring inequity, you the human being spent all that time, while the combined CPUs spent five seconds on your task."

He suggested, in this case, the Web, or software being used, should understand an individual's preferences ahead of time, like, a preference for sunny climates, site-seeing options but not too many museums, etc.

How do we get to these new tools and approaches? Raghavan said it's important the industry remember what spawned the explosive growth of the Internet. "How it came about was openness. The system wouldn't crash just because someone put up a bad HTML page. That principle has worked and openness must persist," he said.

He also made a plug for Yahoo's Build Your Own Search Service (BOSS) initiative that basically makes Yahoo's vast search infrastructure available as a platform for developers to innovate on to create new services.

The Web as a 'Giant Brain'

Nova Spivack, founder and CEO of Semantic Web company Radar Networks, said he's been seeing a shift he describes as the World Wide Web becoming the Wide World Web. "With things like geo-tagging and flexible displays that bring Web access outside the office or home, we're seeing augmented reality become reality," he said. "The Web is becoming more contextually aware."

Spivack, who served as moderator of the panel, said the Web is moving from cloud computing to a more pervasive "ground computing" and asked if this trend is leading the Web to become more like a "giant brain" that thinks and extends human minds. "And if so, is that giant brain going to be Google?"

Peter Norvig, director of research at Google (NASDAQ: GOOG), scoffed at the idea. "I think we already have human brains. We don't want to limit the Web to what a brain can do, but augment what a brain can't do."

Norvig conceded that currently users have to do most of the thinking to get what they want from the Web. "I think we have to do a better job of having a conversation with the user," he said. Google advanced that idea a bit years ago when it introduced the 'Did you mean...?' feature in search results where the search engine suggests the correct spelling for commonly misspelled entries.

"But we need to go much farther to both where there is a clear intent and in cases where it's more exploratory," he said. As an example, he noted someone seeking help for a medical problem might start out a search by entering a symptom like "funny red splotch" which would lead to many ambiguous results and repeated queries. "We have to accelerate the process and not make the user do all the work," he said.

Next page: The evolving human brain