The opening panel was ably moderated by Gary Stein of Jupiter Research. Panelists included executives from a nice mix of Web search engine companies that also operate in specialized categories, such as enterprise search, vertical search, and clustering. We'll stay away from mentioning personalized search, which was declared a dead-end by Vivisimo's CEO, Raul Valdes-Peres. Other panelists include Stephen Baker from FAST, Dave Mandelbrot of Yahoo! and Rich Skrenta of Topix.net.

While there was general agreement that search is fundamental and has become the preferred method for finding information on the Web, this panel also represented the view that search is becoming the preferred first step in finding information within the enterprise and on one's own desktop. Nonetheless, this session also suggested that search has a long way to go in its evolution to provide value to a wider spectrum of users, content-types and specific applications. Vertical search, the current catch-all phrase for all of the above, was a hot topic. Additional methods for discerning relevance beyond links, whether it be clustering, filtering, or trusted metatagging was also a prominent theme. Note my use of the modifier "trusted" before metatagging. Raul cautioned that metatags add little value to textual content, yet Gary pointed out how structured content helps search engines understand meaning.

Clearly progress has been made in the past year or two in how the major search engines utilize content structure and tagging to understand relevance. But there is much progress to be made in how these companies use publisher-created structure/tagging, as well as user-created tagging (e.g., folksonomies), to both understand relevance and to aid users in sorting through the oceans of content available through search engines. This opening panel introduced the recurring theme that search is not the be-all and end-all of information retrieval and management.