We’ve begun implementing some of Facebook’s Open Graph on Alikewise, my dating startup. By Dare Obasanjo’s definition, books are our main “social object”.

(The site is based around the idea that a person will associate themselves with any number of books, with the hope that this will inspire notice by other singles.)

Indeed, the Open Graph protocol is a step toward a more semantic web by allowing us to put agreed-upon metadata into our pages. What it doesn’t do is allow us to define relationships among the metadata. Facebook reserves this right for itself.

For example, books obviously have authors. There is no way to define this relationship, as far as I can tell. We can indicate that a page is about a book or about an author, but not both. Ditto actors and movies, or politicians and their voting history.

In other words, one page = one freestanding social object. We need more.

On my site, I’d like to associate a person with any number of books, in a way that is meaningful outside our site. (A person might choose to identify themselves via Facebook, Twitter or any OpenID, for example.)

Then, any third-party site could aggregate a profile of a person based on Alikewise’s information and others. One might Google “people who like Catch-22” and get more meaningful results.

The Open Graph is helpful in that it pushes a semantic web forward, using Facebook’s market power. And via its Like button, Facebook can indeed relate people to social objects. I would like this ability extended beyond Facebook’s firewall.