People have recently begun using the term "social graph" in reference to one’s network of family, friends, and acquaintances along with the data about those individuals and relationships. The word "graph" here is used in the mathematical sense of a set of points together with lines representing relationships among the points. For this purpose "graph" and "network" are essentially equivalent terms. The Internet is a graph (a set of computers and the communications circuits that interconnect them). So is the Web (a set of documents and the links that interconnect them).
Tim points out that the Internet is an abstraction that hides all of the complexity of the physical communications links. It’s computer-centric in that any computer can talk to any other computer on the Internet without having to know the details of the communications links connecting them.
Similarly, the Web is an abstraction that hides all of the complexity of computers, folders (directories), and the documents in those folders. It’s document-centric in the sense that any document can link to any other document and users can navigate from document to document without having to know the details of the computers and folder structures that hold them. The Web sits on top of the Internet and simplifies the process of accessing information in documents. While the Internet links computers, the Web links documents.
Now, Tim says, along comes another layer in the abstraction hierarchy. The Semantic Web asserts that it isn’t necessarily the documents that are important but the subject matter they cover, the information they convey. The Semantic Web is object-orientation for the WWW. It presents information in such a way that software can reuse it, combine it, and mash it up in ways perhaps never envisioned by the original publisher of the information. It’s a programmable web that hides the complexity of which information is in which document. It’s information-centric in that users can navigate related information without concern for which document holds it.
Tim suggests that when it comes to one’s social graph, what we really need is a way to create, store, publish, and reuse information about ourselves and our social network such that it no longer matters which document (or service or platform) holds which information. Such a structure would constitute a Giant Global Graph – an abstraction that sits on top of the WWW and aggregates all of the interesting data about ourselves and our relationships into a database usable by computers.
This powerful idea shows the flaw in Chuck Russell’s (my fellow Appian Way blogger) analysis of Facebook. (Yes, we’re having the argument here in the blogosphere for everyone to read! As a political science graduate, Chuck LOVES to argue in public. Who am I to deprive him of such joy?) The walled garden may persist for a time, but eventually it gets replaced by a system with radical interoperability. "Small pieces, loosely joined" (as Dave Winer would say) results in a richly diverse ecosystem that over time benefits everyone involved.
For example, AOL used to be an online service – a 1980s era computer bulletin board updated with a graphical interface. Once the Internet and Web became popular, AOL connected to them under the hood but still presented users with a proprietary user interface and search system (remember AOL keywords?). So the question for Chuck is, do you want to bet on Facebook (a walled garden like the AOL of old) or the entire rest of the Internet? Short term, no question Facebook is popular as was AOL in its day. Long term, I’m betting on the Giant Global Graph.
None of this is to say that we shouldn’t work with Facebook for right now since it’s popular. Nor am I saying that Facebook won’t eventually open up and become part of a highly interoperable GGG. I’m just saying that their current model is flawed and our investment in Facebook-specific strategies should be modulated accordingly.
By the way, Tim Bray agrees with me.