The Teachable Search Engine
from the give-it-some-time dept
David Plotnikoff from the San Jose Merc has a column today saying that he wants a more personalized search engine. He wants a search engine that, like Amazon’s recommendations engine, learns from what he clicks on, and guesses at what he’ll find interesting. I’ve actually seen a few research projects that touch on this sort of ability, and I think it’s certainly possible to create such a thing. However, it might be quite a while before it would work well enough to be useful. The column also doesn’t address the issue of what do you do when you want the “other” results? What if yesterday I’m interested in researching sequoia trees, and the next day I want to about Sequoia the VC firm or any other such example?
Comments on “The Teachable Search Engine”
Dynamic
And Amazon’s suggestions are based on content that doesnt change (books, music, dvd’s, etc), whereas a website can drastically change putting a strain on the engine to keep up with its content.
And you cant have a user rating system to judge websites, because that opens a whole new door for people to overrate their sites to get them high on the lists.
Good idea in theory, but it seems it would be extremely difficult in practice.