They have to get a court order at the least, and several governments in south america have asked for the info from google and google, very nicely, told them to go f*ck themselves.
China has got small sets of data from them, but nothing on the scale of yahoo and aol has handed over... yet.
@registering, no browsers identify themselves with a http header that identifies their 'agent'. So when it see's agent as Googlebot it identifies them as a bot. it then turns off things like sessions, login checks, etc so its easier on the site. you can turn it off and it will just appear as another unregistered 'leech'
, however its not as efficient.
Basically, google and other search engines, for better or for worse, have clusters of machines that 'spider' sites... that is go through, consume lots of bandwidth and index an entire site by virtually 'clicking' every link in and out. It does things like calculate a site ranking based on content hits and incoming and outgoing links.
So the best thing you can do for dead-donkey.com is stick it in your signature at other sites, because when they are indexed it will see a link to dead-donkey and say its a better site based on the number of links.