The world wide web has a huge resource of information available to tap. But precisely because of its size, it is also extremely hard to search for the relevant information. Now scientists have found that the web spontaneously organises itself into communities and a new search algorithm could speed up the hunt for the relevant material substantially. Researchers from Princeton-based NEC Research Institute have found that the pages on the web congregate into social groups that focus most of their attention on each other. The new search algorithm ignores the text of the page and only focuses on the link to other sites and so on till it identifies islands with things in common. When the new algorithm was tested it proved to be better than conventional engines like Google and Yahoo.