Open Graph is confusing as hell.
If you watched the F8 keynote, you know that you can like arbitrary things on different websites, provided they’re tagged properly with metadata. That essentially turns that page into a Facebook page behind the scenes (and will be indexed by Facebook search).
So let’s say I like Koyaanisqatsi on IMDB. What Facebook page does that equate to? It equates to Koyaanisqatsi (1982), a page on IMDB (it has no visible “page” on Facebook), which is liked by eight people. Let’s pretend that tomorrow morning Rotten Tomatoes suddenly adds Facebook Like buttons to all its movies. If I go to Koyaanisqatsi on that site, I won’t be marked as “liking” the movie. There’s a namespacing issue going on.
Each of these pages would be considered different entities as far as Facebook is concerned, with different owners, meaning that any feed updates sent by the Koyaanisqatsi IMDB page are not going to be the same ones as the ones from the Koyaanisqatsi Rotten Tomatoes page. If I wanted to get feed updates from both sources, I’d have to become a fan of both pages, meaning that Koyaanisqatsi would likely appear twice in my auto-populated “Favorite Movies” part of my profile.
And let’s not forget that if the people behind Koyaanisqatsi actually wanted to have a Facebook page to send updates out through, they’d be competing for “likes” with the IMDB and (hypothetical) Rotten Tomatoes page.
The Open Graph Protocol’s goal is to “represent profiles of real-world things”. If that’s the case, why can’t I like Koyaanisqatsi the movie instead of liking a page describing Koyaanisqatsi?
One could argue there is already duplication in Facebook pages right now. While that is true, this certainly isn’t helping.