Networks -- Part 2

Here is a little more on the Science of Networks. These points stuck with me:

The Internet is not randomly formed. If it was, most nodes, or websites, would have a similar number of connections to other nodes. But the Internet has connectors, a relatively small number of nodes that have a large number of connections to other nodes. Most nodes have few connections, but these few have many. This is often described as the 80:20 rule. 80% of the links are held by 20% of the nodes.

When networks form, new nodes are constantly added. As they make connections, they have a preference for nodes that have the most connections. The rich get richer... In a social setting these are the popular people in the room. There is a "fitness" factor for nodes that can result is a new node rapidly accumulating connections, and even usurp the connector status of existing nodes. The ideas that the web is a great democracy, where any website can complete with any other is simply not true.

Networks can be directional, where the connections are one-way only. This is exhibited on the Internet. My website may link to your website, but yours probably doesn't link to mine. This affects the software robots used by the search engines (google, etc.) to index websites. They end up only being able to index about half the web. These robots start at a connector website, somewhere in the middle of things. They can swim downstream, but can't get upstream.

These revelations only really started to be understood in 1999. And our understanding of how the Internet forms, grows, and changes over time is still evolving (as is the web.) Once these principals were see in relation to the Internet it started to become clear that other networks, from cell structures to economic markets exhibited similar behaviors.

A great read -- well worth the time, and the audio book is well done.

Albert-László Barabási's book, Linked: The New Science of Networks