I always had this idea of a distributed search engine that uses free CPU time and bandwidth on client machines for crawling and indexing. The code could either

I always had this idea of a distributed search engine that uses free CPU time and bandwidth on client machines for crawling and indexing. The code could either be distributed by <script> tags or as browser plugins The problem is however not so much if it's technically feasible but why users should contribute their CPU time and bandwidth in the first place (particularly on mobile where both are scarce). I suppose enterprise environments might be more amenable to such a system than consumer applications, particularly because most office PCs don't use most of their CPU time anyway. The question is what exactly to do with distributed computing capabilities in such environments.