outfoxing the web
A year and two days ago I wrote about Google's personalized searching, where Google fine-tunes search results based on a profile of your interests. Yesterday I wrote about the platypus extension for greasemonkey, which lets you remix the contents of webpages to your preferences. I think these are both great technologies that reclaim some of the dynamic nature of the web, enabling you to choose what you see instead of passively absorbing everything as though your computer is just a fancy television. (We can thank AOL for the confusion; they marketed the web as though it was a new TV channel and for millions of people that was their first introduction to the net. A friend asked me the other day "what ever happened to chat rooms?" and I had to think about whether I should bother going into the details of IRC and MUDs and MMORPGs when hand-waving would probably work just as well.)
Outfoxed adds a social dimension to content filtering. You and your trusted (Firefox-using) friends share ratings - this page is good, that page is bad, this other one is dangerous - and these ratings are shown in search results. The important thing is that the overall rating isn't just a measure of what your friends think of a site - it is dependent also on what you think of their opinion, i.e. you get to rate the trustworthiness of your friends ratings.
The project is the implementation of Stan James' masters thesis; he hit all the big buzzwords with this one: meta tags, FOAF communities, and personal search optimization. He even accounted for the possibility that some people won't want to do the legwork of tagging sites: you can import publicly posted ratings, which means that (for instance) if you want to see the web the way Lawrence Lessig sees it, you could point your browser at his ratings list and import it into your own. But given the popularity of sites like Flickr, I don't think sparse tagging is going to be a problem.
There are two caveats that I think are important; one is discussed minimally on the site, the other not at all.
First, regardless of how much you like your friends the odds are very high that at least one of them is a total fucking idiot. Lovable, sure, but maybe not so bright. It's an occupational hazard of being human. If you've got a friend that visits sites that install malware, the problem will sort itself out quickly enough as people learn to not trust that person's recommendations, but you could be the one that gets burned in the interim. Given the recent security holes (on every platform, Macs just got hit with a doozy the day after Tiger was released) that allow Bad Things to happen when you visit websites, this could be a serious mess. On the other hand it's not all that different from every interaction you have with other people... you never know who has cooties.
The second problem is one I addressed in the post from a year ago: limiting all of your searches to things you are already interested in seems like a good way to restrict growth. I'm online a lot because I like learning new things; if every site I went to only reinforced things I already knew, I'd be learning nothing. And groups of people who implicitly restrict their viewpoint to that of other group members are called 'cults', not 'online communities', and individual members are called 'ignorant'. (Except recently there has been a shift, and sometimes those groups are called 'neocons', and induhvidual members might be call 'president'.)
Outfoxed adds a social dimension to content filtering. You and your trusted (Firefox-using) friends share ratings - this page is good, that page is bad, this other one is dangerous - and these ratings are shown in search results. The important thing is that the overall rating isn't just a measure of what your friends think of a site - it is dependent also on what you think of their opinion, i.e. you get to rate the trustworthiness of your friends ratings.
The project is the implementation of Stan James' masters thesis; he hit all the big buzzwords with this one: meta tags, FOAF communities, and personal search optimization. He even accounted for the possibility that some people won't want to do the legwork of tagging sites: you can import publicly posted ratings, which means that (for instance) if you want to see the web the way Lawrence Lessig sees it, you could point your browser at his ratings list and import it into your own. But given the popularity of sites like Flickr, I don't think sparse tagging is going to be a problem.
There are two caveats that I think are important; one is discussed minimally on the site, the other not at all.
First, regardless of how much you like your friends the odds are very high that at least one of them is a total fucking idiot. Lovable, sure, but maybe not so bright. It's an occupational hazard of being human. If you've got a friend that visits sites that install malware, the problem will sort itself out quickly enough as people learn to not trust that person's recommendations, but you could be the one that gets burned in the interim. Given the recent security holes (on every platform, Macs just got hit with a doozy the day after Tiger was released) that allow Bad Things to happen when you visit websites, this could be a serious mess. On the other hand it's not all that different from every interaction you have with other people... you never know who has cooties.
The second problem is one I addressed in the post from a year ago: limiting all of your searches to things you are already interested in seems like a good way to restrict growth. I'm online a lot because I like learning new things; if every site I went to only reinforced things I already knew, I'd be learning nothing. And groups of people who implicitly restrict their viewpoint to that of other group members are called 'cults', not 'online communities', and individual members are called 'ignorant'. (Except recently there has been a shift, and sometimes those groups are called 'neocons', and induhvidual members might be call 'president'.)
1 Comments:
-stan
return to front page