I find it quite amazing that people, when they see the profile page of "John Doe" on Facebook, assume that the page is indeed controlled by John Doe and allows them to contact John Doe. (No, in fact, they wouldn't do that for someone called "John Doe", but let's assume that the name isn't "John Doe" but something sufficiently unique to ensure that there aren't multiple people with that name.)

If you think about it, this is absurd; what they are seeing is a page managed by Facebook, which pretends that they gave the control of the page to someone who claimed to be John Doe. Even if we assume that Facebook isn't interfering, at no point whatsoever did anybody check that the guy is indeed John Doe.

In practice, things work out quite well because there aren't that many people who are willing to take the time to impersonate someone else on Facebook. Nevertheless, I know that some people had to complain to Facebook because they had some sort of enemy who created an (insulting) Facebook profile with their name. This shows that there is a problem; theoretically, when confronted which such a page, nobody should assume that the page has anything to do with the person whose name appears on the page...

(Of course, this problem isn't restricted to Facebook, but extends to most of what you find online. Not everything, though (notable exceptions are the OpenPGP web of trust, and the HTTPS certificate authority system which is itself quite flawed).)