Right, without joining the "is napster bad?" debate, its problem was (as far as I know) that it had a central server that indexed all the users and mp3s. This meant that it was possible for files to be blocked at the record companies' request. Mis-spelling band and/or song names will get round a lot of the blocks - thing is, because its users have diminshed, so have the amount of available mp3s. The peer-to-peer applications mentioned in the 3rd post (e.g. gnutella) do not index using a central server, instead they seek out other users online and link to yet more. Which means more files, less blocks.