A reading from last week, by the same author who wrote the article on the 80/20 rule (in fact, they are different chapters from the same book) discusses how these Power Laws affect different networks and how free-scale networks begin and grow.

It talked about the way in which unplanned networks (in the sense that they don’t have a premeditated structure, like a building might) inevitably create these hubs, from which most other parts of the network are connected. These hubs are the same thing as the spikes in the Power Law graphs.

These hubs are usually the oldest parts of the network, as they have more time to develop more links to other parts of the network. For example, Hollywood film producers who work for big companies will inevitably know a lot of people, who have less ties than they do.

This raises another interesting point. That hubs have many many links, but theses links are weak. Whereas parts of the network that are less linked, generally have stronger links. It ties in with what was discussed in the lecture, “You don’t get jobs from friends, you get jobs from acquaintances”. Those loose links are generally stronger, and the tighter links are generally weaker.

Generally, the growth of a node can be determined by two factors.

“A. Growth: For each given period of time we add a new node to the network. This step underscores the fact that networks are assembled one node at a time.
B. Preferential attachment: We assume that each new node connects to the existing nodes with two links. The probability that it will choose a given node is proportional to the number of links the chosen node has. That is, given the choice between two nodes, one with twice as many links as the other, it is twice as likely that the new node will connect t0 the more connected node.”

But the thing that I found most interesting with the article was the thought it left us on (especially in regards to my comment in my last post about adjusting something so that it better fits your expertise), about the next step for scientific research into free-scale network models.