What is Linked Data? ...featuring Yuri the unicorn

I did an IAP on LinkedData at MIT and one of the session's was given by Sir Tim Berners Lee, creator of Internet and now creator of Semantic Web.

LinkedData has amazing power and capability,the only challenge is adopting it. NYTIMES has all its articles on Linkeddata, DBPedia is an RDF version of Wikipedia.

On the other side all the companies are opening their API's, linked data has a similar vision of data mashups. These two overlaps in few aspects.

http://www.ted.com/talks/tim_berners_lee_on_the_next_web.htm...

Can someone tell me how this isn't a doomed effort to impose a universal ontology on the web?

http://en.wikipedia.org/wiki/Ontology_(information_science)

Such an effort is doomed because in the actual "real world", there isn't agreement on what exact "real world objects" exist and so we can't simply create a single hierarchy of the world's data akin to the Dewy-Decimal system for organizing books.

"Yuri the unicorn" is actually a good example. If such a unicorn "exists", he is likely a character in cartoon or comic book. If he's popular, information about him might meaningful and important. Who provides it? The copyright holder? Wikipedia? Whoever registers the URL first? (and don't even start on the question of "does he really exist?")

The web has struggled-with and provided a half-way solution to these problems with the combination search-engine and leaving the decision up to the user. But an effort to create a single authoritative hierarchy of "real world things" will fail and what's a "link to a real-world thing" if it's not authoritative aside from a needless layer of indirection?

Edit: The web's "half-way solution" is actually really good and amazing all considered and the "semantic" efforts to go "forward" from this in many ways seem to be steps backward.

I wish some major player (e.g. Google) started backing linked data. It would speed up adoption tremendously.