I was thinking about the recent lawsuits in the US and dreamt up this solution:
Currently the record industry is tracking kazaa users based on queries for specific file names. Considering the kazaa network is p2p with no centralization, it is feasable to simply use the record industries ideas against them.
Considering the record industry is now searching for specific mp3 files out there as a way to target kazaa users we can only assume this is because these specific mp3's are traded more often than others. As such it is assured that a search on the kazaa network for these "specific" files will result in mp3 traders with many files. The thinking here is (if you looked at the list, it had the beattles and avril lavigne) ppl having one or more files on the "list" are users that potentially have a large collection of illegal mp3's due to the large variation of music genres.
What I am getting at here is as a user connects to the kazaa network with highly common files being shared on his/her system the kazaa application automatically proxies the shared file list to another kazaa user. Kind of a self propagated, proxied share database. It can even go as far as proxying downloads via a method demonstrated in apps like Download accelerator. In this scenario, someone searches for ArtistA SongA. One of the many search results displayed will point to a specific kazaa user hosting the mp3...but this kazaa user isn't acttually hosting the mp3, it is hosted by another kazaa user. When that someone goes to download an mp3 they make the request to the proxy, the proxy then in turn does a search of the kazaa network and instructs other kazaa machines to initiate connections through open proxies also displayed as search results in the person's search.
In effect this idea would make p2p file sharing one large logical internetwork where every node is proxied through other nodes. Like one huge botnet. This probably would require a whole new p2p filesharing app/network but it does certainly make the recording industries life harder with significantly more overhead to determine who is actually sharing what and how much they are sharing.