I’m working on a podcast…. stay tuned for an episode, hopefully by the end of the week.
Archive for February, 2005
I still have 50 Gmail invites, I’m also now an Orkut member, so I can get you invites for there too. Let me know if you’d like one of either.
I just opened Google Maps a second ago (because I learned that it now support Safari) and noticed something interesting. I you open without an addresss (and get the view of the whole us), then zoom in, you’ll see that its centered n Coffeyville, Kansas, which happens to be the location of one of Amazon’s major operations centers.
I wonder if:
1. Amazon and Google each calculated the center-most town in America, or
2. Google decided to follow Amazon, or
3. this is just, random.
So, I’ve been working on a school research project which has morphed into a personalized search research project. The ideas I have about it are somewhat rough and my explanations of it often include:
me: “…and then we’ll do something here and come out with what we want.”
them: “How are you going to do that?”
me: “I don’t know.”
them: “Good luck.”
me: “But once we figure it out, we’ll grow fast and them buy out Google, that way we don’t have to use their webservices, we can have direct access to their data.” :-)
The problem we’re tackling is immensly hard in both technological and a human-design ways- I don’t claim that we can create a search engine that can beat Google, but I think the type of search engine we’re looking at has an important quality which will be neccessary beat Google: personalization.
Let me explain why I think personalization is so important:
A long, long time ago, there was no Google. What we had were all text-based search engines. These search engines used only the content of pages in order to do their search and, more importantly, ranking.
Then along come some people name John Kleinberg (at Cornell) and Sergy Brin and Larry Page (at Stanford). Both groups, it appears, came to similar ideas concurrently (Newton/Leibniz style). Their idea was simple: make use of the link structure of the web as a way to measure the relevance of pages. Kleinberg and Brin/Page dealt with the details differently (and with varying levels of success, Kleinberg’s still at Cornell, Brin and Page run Google).
So, the point I’m trying to make here was that Google was a paradigm shift in the way search engines and in what kind of data they make use of in performing their searches. In order to beat them, there will have to be another, equally drastic shift- a shift away from a universal ranking of relevance towards personal ones.
This my friends, is personalized search. And people have been working on it for a decade with very little success. Do I think I can do better? Maybe. Anyway, its worth a shot, eh?