Wikipedia founder Jimmy Wales thinks the masses can build a search engine that offers what Google does not.
"For developers, it would be kind of cool to say 'Google has 11 gazillion dollars, and we're going to get a bunch of people together on the Internet, and we're going to kick their butts,' " he said. "That's cool, right? I think it's cool."
Wales spoke publicly for the first time, at New York University last week, about his vision for an open source, for-profit Search Wikia. He said Monday that he believes the first beta will launch within a few months.
Wales described himself as a "free software religious zealot" and made it clear that his faith in free software, democracy, community, and transparency will drive the development of Search Wikia.
"Search is a fundamental part of the infrastructure of the Internet and therefore it is a fundamental part of culture and human society as a whole," Wales told a crowd hosted by Free Culture at NYU.
He said search should be open, transparent, participatory, and democratic. He said Search Wikia would be free, "as in speech, not as in beer." Though Wales said, "I love Google," at least three times during his talk at NYU, he hopes to create something more meaningful, more transparent, and ultimately better than Google.
"People want to know how these searches are being determined," he said. "I love Google. I love Yahoo. I don't think they're doing anything evil."
Wales said he also loves the court system because people can go in and watch the proceedings. He wants to deliver the same transparency to search. Contributors will publish, test, research, and modify the algorithms.
Beginning with Lucene and Nutch (two A "trust network" of users, similar to those who contribute to Wikipedia, will discuss and debate what should come next. As search heads toward greater personalization, Wales believes people can come up with smarter, more relevant searches than machines. He indicated a belief that collective wisdom will improve search more than knowing everything about an Internet user.
He characterized his description of the search plans as a "raw brain dump" and added that he has not figured many of the details, including how the project will protect privacy.
Though spammers will likely try to skew the system in their favor, Wales said the project would allow the "trust network" to overwhelm spammers, the same way people swarm around those who violate rules on Craigslist and Wikipedia.
"Flickr has almost no spam," he said. "When I first saw it, I thought there was going to be acres and acres of porn spam."
Wales said it is a mistake to obscure things for security.
"If you're relying on people not knowing how the system works, you've got a big problem," he said.
Already, Wales, who described himself in a recent Newsweek article as "pathologically optimistic," has an international following. Wikipedia, the not-for-profit venture he led, boasts more than 2 million articles in about 250 languages. While the 40-year-old says he still views himself as "that guy who was sitting around at home in his pajamas typing on the Internet," he has financial backing from venture capitalists and Amazon.com.
Wales said the search will use a traditional ad revenue model and rely on some hired staff but he declined to be more specific.
"It has to be fun," he said. "If it's not fun, people are not going to do it."
Still, don't expect any big launch parties when Search Wikia comes out in beta.
"As soon as we announce beta, people will chime in, 'It's going to suck,' " he said. "In order to get the media to calm down, we're going to say, 'We know it sucks. Just come try it. If you don't like it, come back in a year.' "
Google on Wikia search
Google (goog) - the greatest search engine in the galaxy - reports next week. Ahead of that report, Yahoo said that it's going to rank its results in a similar fashion to the way Google ranks them, theoretically resulting in higher click-through rates for Yahoo (yhoo).
Meanwhile, a number of search engines are trying to be Google and Yahoo killers, including Wikia search, Jimmy Wales's new endeavor.
Wales, the founder of Wikipedia, is going to try to do to the big search giants - Google (goog) and Yahoo (yhoo) -- what his Wikipedia venture did to Britannica - make it somewhat less relevant.
So, what did Google's director of research. Peter Norvig, have to say about Wales' pursuit? Here's our interview via email.
Me: What are the flaws of a wiki or community-driven search? And, what could be the advantages?
Norvig: Judgments from users are certainly important for search. In a way, hyperlinks (the basis of the Google PageRank metric) are user judgements, but the users are limited to webmasters. The more the merrier, so it is good to get feedback from others as well. Google and other search engines do use user feedback, mostly implicit feedback in the form of clicks, with some explicit feedback. We experimented with smiley/frowny faces in the toolbar, and with various types of feedback on search results. Currently we have a " Dissatisfied? Help us improve" link at the bottom of the page; at times we've asked for positive feedback as well. There are two main challenges for community-driven search.(1) Long tail. While there are many popular queries (See http://www.google.com/press/zeitgeist_monthly.html ) there are also millions of queries per day that we've never seen before. (2) Spam. The more you get away from an academic setting and into one where there is money to be made, the more you'll have spam, deception, people trying to game the system. It seems that the best approach is an algorithmic system that uses community-driven data when it is available and deemed reliable, and does not use it when it is not.Me: Is SearchMash a Google wiki-styled search engine? If so, what is the goal of this new product?Norvig: That's our playground for trying out new ideas -- of all kinds -- in a non-branded search engine where we are more free to experiment with different algorithms and user interfaces. It is useful because we don't have to adhere to the user's expectations for what Google is. It does include " Feedback: Were these results useful to you? ..." but it is not exclusively social/wiki-style; it is for anything we want to try out.
Me: So, are you testing out a wiki-styled search engine?
Norvig: Feedback from users is important, and we will continue to use the sources of user feedback we have been using, and will experiment with more in the future. I think that Google Co-op Custom Search Engine may be the largest current collection of user-generated information for search. http://google.com/coop/cse/Google on Wikia search
Google (goog) - the greatest search engine in the galaxy - reports next week. Ahead of that report, Yahoo said that it's going to rank its results in a similar fashion to the way Google ranks them, theoretically resulting in higher click-through rates for Yahoo (yhoo).
Meanwhile, a number of search engines are trying to be Google and Yahoo killers, including Wikia search, Jimmy Wales's new endeavor.
Wales, the founder of Wikipedia, is going to try to do to the big search giants - Google (goog) and Yahoo (yhoo) -- what his Wikipedia venture did to Britannica - make it somewhat less relevant.
So, what did Google's director of research. Peter Norvig, have to say about Wales' pursuit? Here's our interview via email.
Me: What are the flaws of a wiki or community-driven search? And, what could be the advantages?
Norvig: Judgments from users are certainly important for search. In a way, hyperlinks (the basis of the Google PageRank metric) are user judgements, but the users are limited to webmasters. The more the merrier, so it is good to get feedback from others as well. Google and other search engines do use user feedback, mostly implicit feedback in the form of clicks, with some explicit feedback. We experimented with smiley/frowny faces in the toolbar, and with various types of feedback on search results. Currently we have a " Dissatisfied? Help us improve" link at the bottom of the page; at times we've asked for positive feedback as well. There are two main challenges for community-driven search.(1) Long tail. While there are many popular queries (See http://www.google.com/press/zeitgeist_monthly.html ) there are also millions of queries per day that we've never seen before. (2) Spam. The more you get away from an academic setting and into one where there is money to be made, the more you'll have spam, deception, people trying to game the system. It seems that the best approach is an algorithmic system that uses community-driven data when it is available and deemed reliable, and does not use it when it is not.Me: Is SearchMash a Google wiki-styled search engine? If so, what is the goal of this new product?Norvig: That's our playground for trying out new ideas -- of all kinds -- in a non-branded search engine where we are more free to experiment with different algorithms and user interfaces. It is useful because we don't have to adhere to the user's expectations for what Google is. It does include " Feedback: Were these results useful to you? ..." but it is not exclusively social/wiki-style; it is for anything we want to try out.
Me: So, are you testing out a wiki-styled search engine?
Norvig: Feedback from users is important, and we will continue to use the sources of user feedback we have been using, and will experiment with more in the future. I think that Google Co-op Custom Search Engine may be the largest current collection of user-generated information for search. http://google.com/coop/cse/
Meanwhile, a number of search engines are trying to be Google and Yahoo killers, including Wikia search, Jimmy Wales's new endeavor.
Wales, the founder of Wikipedia, is going to try to do to the big search giants - Google (goog) and Yahoo (yhoo) -- what his Wikipedia venture did to Britannica - make it somewhat less relevant.
So, what did Google's director of research. Peter Norvig, have to say about Wales' pursuit? Here's our interview via email.
Me: What are the flaws of a wiki or community-driven search? And, what could be the advantages?
Norvig: Judgments from users are certainly important for search. In a way, hyperlinks (the basis of the Google PageRank metric) are user judgements, but the users are limited to webmasters. The more the merrier, so it is good to get feedback from others as well. Google and other search engines do use user feedback, mostly implicit feedback in the form of clicks, with some explicit feedback. We experimented with smiley/frowny faces in the toolbar, and with various types of feedback on search results. Currently we have a " Dissatisfied? Help us improve" link at the bottom of the page; at times we've asked for positive feedback as well. There are two main challenges for community-driven search.(1) Long tail. While there are many popular queries (See http://www.google.com/press/zeitgeist_monthly.html ) there are also millions of queries per day that we've never seen before. (2) Spam. The more you get away from an academic setting and into one where there is money to be made, the more you'll have spam, deception, people trying to game the system. It seems that the best approach is an algorithmic system that uses community-driven data when it is available and deemed reliable, and does not use it when it is not.Me: Is SearchMash a Google wiki-styled search engine? If so, what is the goal of this new product?Norvig: That's our playground for trying out new ideas -- of all kinds -- in a non-branded search engine where we are more free to experiment with different algorithms and user interfaces. It is useful because we don't have to adhere to the user's expectations for what Google is. It does include " Feedback: Were these results useful to you? ..." but it is not exclusively social/wiki-style; it is for anything we want to try out.
Me: So, are you testing out a wiki-styled search engine?
Norvig: Feedback from users is important, and we will continue to use the sources of user feedback we have been using, and will experiment with more in the future. I think that Google Co-op Custom Search Engine may be the largest current collection of user-generated information for search. http://google.com/coop/cse/Google on Wikia search
Google (goog) - the greatest search engine in the galaxy - reports next week. Ahead of that report, Yahoo said that it's going to rank its results in a similar fashion to the way Google ranks them, theoretically resulting in higher click-through rates for Yahoo (yhoo).
Meanwhile, a number of search engines are trying to be Google and Yahoo killers, including Wikia search, Jimmy Wales's new endeavor.
Wales, the founder of Wikipedia, is going to try to do to the big search giants - Google (goog) and Yahoo (yhoo) -- what his Wikipedia venture did to Britannica - make it somewhat less relevant.
So, what did Google's director of research. Peter Norvig, have to say about Wales' pursuit? Here's our interview via email.
Me: What are the flaws of a wiki or community-driven search? And, what could be the advantages?
Norvig: Judgments from users are certainly important for search. In a way, hyperlinks (the basis of the Google PageRank metric) are user judgements, but the users are limited to webmasters. The more the merrier, so it is good to get feedback from others as well. Google and other search engines do use user feedback, mostly implicit feedback in the form of clicks, with some explicit feedback. We experimented with smiley/frowny faces in the toolbar, and with various types of feedback on search results. Currently we have a " Dissatisfied? Help us improve" link at the bottom of the page; at times we've asked for positive feedback as well. There are two main challenges for community-driven search.(1) Long tail. While there are many popular queries (See http://www.google.com/press/zeitgeist_monthly.html ) there are also millions of queries per day that we've never seen before. (2) Spam. The more you get away from an academic setting and into one where there is money to be made, the more you'll have spam, deception, people trying to game the system. It seems that the best approach is an algorithmic system that uses community-driven data when it is available and deemed reliable, and does not use it when it is not.Me: Is SearchMash a Google wiki-styled search engine? If so, what is the goal of this new product?Norvig: That's our playground for trying out new ideas -- of all kinds -- in a non-branded search engine where we are more free to experiment with different algorithms and user interfaces. It is useful because we don't have to adhere to the user's expectations for what Google is. It does include " Feedback: Were these results useful to you? ..." but it is not exclusively social/wiki-style; it is for anything we want to try out.
Me: So, are you testing out a wiki-styled search engine?
Norvig: Feedback from users is important, and we will continue to use the sources of user feedback we have been using, and will experiment with more in the future. I think that Google Co-op Custom Search Engine may be the largest current collection of user-generated information for search. http://google.com/coop/cse/
Subscribe to:
Posts (Atom)