Why the Graph and the Semantic Web are not the same

A few weeks ago, Tim Berners Lee, the inventor of the Web, wrote a blogpost on The next level of the internet. The main idea is the following:

• the internet was the first step: a network of computers
• the web was the second step: a network of documents
• the graph will be the third step: a network of relationships between people and data

Although the concept of the Social Graph has been touted since long by various Web 2.0 bloggers, it is remarkable to see that Tim Berners Lee is adopting this idea. For those who don't follow the evolution of the internet, Tim has always been an advocate for the concept of "The Semantic Web".

Most people now interpret the idea of "The Graph" as a synonym for "Semantic Web". To me, both concepts are quite contradictory.

The movement of the semantic web has had a hard time, since they are trying to advance the benefits the web brings to us by advocating that we write computer-readable code instead of just human-readable code. The reasoning is that is you put e.g. a street address in HTML, only humans can detect this as an address. But computers will just see words and not figure out that the words are to be interpreted as a street name. This reasoning brings us to the idea that we should publish our information according to what is called a "taxonomy": a pre-defined structure of information. An example is that we agree that a street address consists of a street name, number, city name, state and country. If we then also agree on how we write this down (e.g. in an XML syntax), computers will be able to retrieve meaningful information from websites and do something with it.

If you think of the fact that the internet became popular in the first place because it allows anyone to express their ideas and creativity by publishing a simple HTML document, you see that this is in conflict with the centralized approach of defining taxonomies. This is why the semantic web has failed so far (after 3 years of continuing efforts) and will never succeed.

The social graph, on the other hand, is about relations between people. The biggest movement we see today in the web is that people start to emphasize the importance of their relations to other people. This is where the success of Myspace, Facebook and Linked-in comes from. If you are active on any of those sites, you'll notice that the information itself becomes less relevant then the links between the people that publish or reference the information.

This is a fundamental concept to social networking: relations between people and information are more important then the information by itself. In other words: the information only gets it's meaning by being tied into a network of connections. This is essentially the same thing that happens with real life information: Your perception of a message is based on how others that you value perceive that message. That is how opinions are formed, guided and manipulated. Just look at the Monica Lewinski affair or the public opinion on the US war against terrorism.

The conclusion of this all is for me that whilst the semantic web tries to capture information based on the significance attributed to it by its author, the social graph tries to capture information based on the significance attributed to it by its readers.

If I select a bunch of text on a web page and add it to my address book, I implicitly indicated that this text is an address. If I drag this address to a restaurant review site and write a comment, I told the system that this is the address for a restaurant. Capturing this user input is much easier then trying to fit all possible data in a predefined taxonomy.

What we need for the social graph to become our frame of reference is platforms: platforms that allow us to reuse information, platforms that put the user in control over what happens with his own data. Citylive is creating such a platform. It's not easy, but that's why we have the best team in the world. If you want to learn more about platforms, you should read the post of Marc Andreessen (founder of Netscape) on the 3 levels of internet platforms.

So Tim Berners Lee didn't just invent a new word for the semantic web by calling it the GGG (Giant Global Graph :)). To me, he actually admitted that the best way to come to a structured web is by using the structures and links that are added by the users and not by the authors.

Interestingly, he makes the valid point that without structure the web will never become successful on mobile devices. Now you know why Citylive as a mobile company puts so much effort in letting its users create their own services and upload them for their friends.