How about a distributed query system?
I have been thinking for some time if a distributed search engine would be feasible. Usually, when people talk about distributed search engines they have in mind something like a cluster of machines where procesing power and other resources are distributed. My proposal, of a Distributed Query System (Draft, v2.0) is more about delegation than distribution.
My approach could possibly solve three of the main problems (that IMOHO) modern search engines have a hard time dealing with: latency, ignorance and web centrism.
Keep in mind this is just a draft, food for thought. I just thought it would be better to make it public than keep thinking about it until it is outdated. Any comments are more than welcome!