I’m working on a search engine crawled by users themselves. If I think about it, I'm not interested in 99% of the web. Therefore I don't need to deal with exhaustively crawling the entire web. Hypothetically, if I could access only the sites visited by the HN crowd, that would be fine by me. Somebody in an art community might feel similarly. If there were some metric the crawlers could use to rate a page (time spent…
Take that idea and evolve it into a federated search service. If your site is in the "fishing" ring then your pages get indexed into the fishing index and the search widget on your site would potentially return results from all the sites in the same ring.
I guess I want SearchRing.