what traditional search engines, including Bing and Google, are currently capable of, he said. This dynamic, intuitive searching requires things that “just aren’t really feasible with keyword syntax,” Hall said.
Still, many struggle to see how search engines can keep up with the need to sort, search, and categorize the mass of multilingual, cross-continental content on the Web. For users who want to see more personalization—search results that integrate comprehensive, real-time, social networking data with vertical, and even emotional, content and user-unique applications like task management, to provide a customized experience—the vision may be far ahead of the reality.
“The technical challenge for that is limited to tracking the user,” Etzioni said. There is plenty of consumer information available to work with, but an issue arises when it comes to the question of personal privacy, he added. “I think a lot of people even in this audience would say they don’t want that.”
The answer to sorting through the innumerable tasks users say they want their search to include, according to Shum, lies in user participation—creating an “overall search ecosystem” full of impassioned users who contribute and determine the “right direction” for the future of search.
Meanwhile, Etzioni emphasized the importance of semantic search. “Most people don’t search for fun. There’s a task in mind,” he said. “And it’s still the case that there are a bunch of tasks—just planning a vacation—that take hours and days to sort through.” He imagines a day “in the not too distant future” when something that now takes hours will take only minutes to search—“when you have an engine that understands the semantics of what’s involved in that task, including what’s being said in reviews.”
When asked the question of what key operational and technical changes search engineers at Microsoft and Google have made in the last decade, the panelists—and Bershad in particular—pointed to developments in comprehensiveness of search results. “Ten years ago, you could get most of the Web on a couple of PCs,” Bershad said. “Search is just as good [now], but the Web is a lot bigger.”
Lastly, the Bing and Google panelists sparred over what is left for each of their companies to solve. “Why should an engineer work at Google? There are not many more problems for them to solve anymore,” Shum said. “We really try hard to look at what Google is good at— comprehensiveness of index, faster response…There are three problems in search: relevance, performance, user experience. A lot of those areas, Google has been leading, and we have been trying to catch up.”
Bershad wouldn’t argue with that, but he challenged Shum’s premise. “If there was nothing left to solve, I wouldn’t have a long list of hard problems on my whiteboard,” he said, reiterating the need to be comprehensive. “For that search to be meaningful, it needs to be meaningful across all the information.” That includes information from around the world, in every language and every medium. “Search needs to work across Web pages, it needs to work across newspapers,” he said. “It needs to work across your medical information, it needs to work across all information in order for it to ultimately be useful.”
