Next-Level Discovery Panel
(Marti Hearst) Discovery - finding things you’ve never seen before.
Search - What’s the Next-Level for Search?
Today: Navigation / Orienting is finding (or not finding , see Search fatigue) what you want one step at a time, each time narrowing in until you (hopefully) find it. - vs. Teleporting. (one jump) Teleporting means that you get to what you want in one or two clicks. The Verticals are very good at this because they have worked at it so hard.
Convergence of Two Trends: Massive collection of implicit use behaviors and better handling of long tail queries and Natural Language Processing (NLP)(Semantic Web)
Note to self: I used to teach Public Speaking at the University of Virginia. It would be good for all speakers to “clock” themselves to see iftheytalktoofast! (Itmakesitreallyhardtotakenotes.)
Lou Paglia - Factiva: They take 12,000 news sources and deliver the “right information to the right person.” Impressive.
Jeremy Miller - Wikia. (Jabber) Open protocols are very important for the Next Level.
Steve Larson - Krugle. (An Alt - yeah!) Indexes are important. Are the Alts like cable TV stations? There were the Big 3 - ABC, NBC, and CBS, and now there are 100 channels, just like the Top 100 Alternative Search Engines.
Bradley Horowitz - Yahoo! - “Interestingness” on Flickr (a Yahoo! company). Apparently there are people out there who are trying to ‘game’ the search results by abusing what they figure out about the underlying algorithms, hmmm. Black hat SEO?
Q: Does Search Suck? Steve - this very war on the engines vs. the gamers has degraded the success of search.
Marti - Search is good because there is so much more information accessible to the engines. But Site Search (Intranets) sometimes sucks.
Lou - Search is good. Enterprise search is not so good. The hidden web remains a challenge.
Jeremie - Search sucks in that it is not living up to it’s potential. Bing!
Bradley - Yahoo! Answers Ah - in Korea people have access to the Internet, but the information was not codified on paper or a webpage - it was literally in people’s heads. So how do you search people’s heads (an interesting part of the invisible/hidden web?). You ask them over the ‘Net. And that’s what Yahoo! Answers (or Hakia’s Meet Others) or similar approaches try to do.
The dark web - information that resides totally within corporate systems. A regular person has no access to that information. Should they - should the data be free? (or “Information wants to be found”)
Can a search engine be devised that would have as the search results the history of a page - like the Wayback Machine perfected as a search engine? Try to image a search for a web site using SpaceTime’s 3-D results. You could just flip, flip, flip backwards in time!
See also www.Trendpedia.com
Overall, most people tend to want the very latest information on a subject. I know that if a Search article is not dated in the past month (Oct/Nov), I probably will not bother to read it.
Coffee break!
November 6th, 2007 at 6:47 pm
While in some cases it’s true search engine quality is degraded by “gamers”, in most cases the results the gamers are putting up are more relevant than others. If they target a keyword, they make SURE that site is suitable to the keyword.
The large scale scraper/cloaker sites though, I will admit degrade the quality, even if they normally only suck up long tail search engine results.