7 ms·
It might interesting to try and design a reverse-search engine, a kind of internet spam filter. Instead of being designed to promote relevant content, it would
by Perceval 17y ago
It might interesting to try and design a reverse-search engine, a kind of internet spam filter. Instead of being designed to promote relevant content, it would be designed to explicitly not return crap content.
Rather than trying to improve Google's resiliency against noise, it might be more fruitful to turn the problem around.
- GHFigs 17y agoThe answer is people. People you trust to not provide you with crap. Sometimes that's friends, sometimes that's family, sometimes that's strangers. It's never social media webcocks who want to "friend you", never bloggers offering you "n Ways to X your Y" for their $5/mo. in AdSense earnings, and only rarely the nebulous "crowd".
- ramchip 17y agoNot so sure. When I'm looking for information on, say, WWII history, I can't ask someone or go on a trusted website like when I'm looking for information on a field I'm familiar with. I don't have any acquaintance or favourite websites already, which is why I'm using a search engine. The large amount of [quiz,farm,friend of the day] spam on my facebook, and the various hoax chain letters my not-too-web-savvy aunt sends us sometimes, also make me a little jaded on the value of 'social $whatever' for information gathering.
- GHFigs 17y agoWWII history, I can't ask someone or go on a trusted website like when I'm looking for information on a field I'm familiar with. WWII history isn't such a hot topic for crapflooding, either. Search is still useful there. You're also still applying trust heuristics based on the results--"battle of the coral sea" returns results from Wikipedia, history.navy.mil, worldwar2history.info, history.sandiego.edu, etc. You probably have some idea about the content of each. Ultimately your decisions about what to click come down to your perception of trust in their authors' ability to provide you the information you desire. I'm just guessing, but I suspect most people here are familiar with the phenomenon of doing a Google search only to end up clicking on the Wikipedia result that invariably makes it into the top few results. This is a sign that Google's PageRank still aligns reasonably well with our own estimation of trust for those search terms. As for social whatevers: that's why I said "People you trust to not provide you with crap." You can't trust most people to do that. Least of all when given a megaphone and/or a financial incentive.
- vaksel 17y agoisn't that kinda what the Yahoo directories were supposed to be? Didn't turn out too well, from what I remember.
- Perceval 17y agoNo no. I don't mean locating good content and putting it in a directory. I mean actively searching out bad content and then making sure that it never appears. Google's approach is to actively search out good content and try to make that rise above the noise. My approach would be to actively search out noise and then make sure that never pops up in search results.
- epi0Bauqu 17y agoI use this strategy at Duck Duck Go. See the Parked Domains Project: http://www.ivegotafang.com/ http://www.ivegotafang.com/
- ThinkWriteMute 17y agoThat's not entirely true. Google actively de-lists non-content pages. Ask any SEO IM.
- Perceval 17y agoI wonder to what degree their de-listing efforts are algorithmically driven (using Page^H^H^H^HSpamRank™), and to what degree they are ad hoc reactions to obvious/reported abuse.
- ThinkWriteMute 17y agoMost of it is actually college kids getting paid per site if I remember correctly.