subscribePaul Graham's blog considered harmful by Google... [false positives and their consequences]
Friday, June 8, 2007
For quite some time now, Google has focused in on-line security. One major goal to achieve is identify potentially harmful sites and report them to the users before they visit them. To do that, a link is displayed in the search results page under suspicious pages saying "this site may harm your computer".
Shelley Powers wonders on her blog:
It's common knowledge that google can destroy (or severely harm!) your web business in seconds just by lowering your PageRank. Now imagine labelling your site as potentially dangerous for the users. Just marking it as harmful could already be a huge damage for your web business. But it doesn't stop there. G almost completely eliminates the possibility a user ignores the warning and visits your site by clicking on Google's result page. To actually visit a site that google thinks wants to harm you, you have to copy and paste the url from google's warning page to the browser's location bar (by the way, a very large number of web users have no idea what a location bar is!)
Two weeks ago, I captured a photo of google's result page identifying Paul Graham's blog at infogami as a site that may harm your computer!!!
This lasted for at least 3 more days after I noticed it. It is now fixed. This is pretty crazy. By examining Paul's blog source code, you can see it's only pure html and text! Text about startups, web businesses, acquisitions, money, life, etc. No weird javascript, annoying pop-ups or redirects.
Google is fine tuning a lot of things daily in their search engine that almost nobody notices. I only noticed this was corrected rather quickly because I was checking it every day.
But I couldn't help but wonder what a false positive would mean if it happened to a small struggling to succeed tech povera start-up.
Shelley Powers wonders on her blog:
Of course, then that begs the question: how accurate is Google's algorithm when we can't, yet again, see what factors it uses on which to base its decision? There's a major difference between a site having low page rank, and marking a site as potentially harmful.
It's common knowledge that google can destroy (or severely harm!) your web business in seconds just by lowering your PageRank. Now imagine labelling your site as potentially dangerous for the users. Just marking it as harmful could already be a huge damage for your web business. But it doesn't stop there. G almost completely eliminates the possibility a user ignores the warning and visits your site by clicking on Google's result page. To actually visit a site that google thinks wants to harm you, you have to copy and paste the url from google's warning page to the browser's location bar (by the way, a very large number of web users have no idea what a location bar is!)
Two weeks ago, I captured a photo of google's result page identifying Paul Graham's blog at infogami as a site that may harm your computer!!!
This lasted for at least 3 more days after I noticed it. It is now fixed. This is pretty crazy. By examining Paul's blog source code, you can see it's only pure html and text! Text about startups, web businesses, acquisitions, money, life, etc. No weird javascript, annoying pop-ups or redirects.
Google is fine tuning a lot of things daily in their search engine that almost nobody notices. I only noticed this was corrected rather quickly because I was checking it every day.
But I couldn't help but wonder what a false positive would mean if it happened to a small struggling to succeed tech povera start-up.
