Skip to main content



Edge types in social networks

One of the salient features of Facebook is the undirected edges connecting users. Information resulting from either user’s actions traverses the edge and makes itself visible upon the neighbor’s news feeds. This uncontrollable flow is not evident in all networking sites. Other sites have information dispersion models which provide greater preservation of privacy, or an improved ability to filter noise. However, I propose a solution which more completely succeeds in this goal.

Twitter is an old player in the social networking scene. Each user maintains an account, represented by a node. Fellow users construct directed edges pointing from these nodes to themselves. The key point is that receiving information is an opt-in process. A message is transmitted to the universe at large, and anyone who opts in may read it.

Google Plus, a newer player, adopts a network structure diametrically opposed to that used by Twitter. Users create so called ‘Circles,’ sets of users to whom they may broadcast messages. The targets of their messages can filter them out, but cannot opt out of receiving them. This suggests that the sender of a message can create a directed edge from himself to the target.

An interesting feature of Google Plus is the directed nature of Circles. A broadcaster attempting to partition a group of contacts has no issues in doing so. A subscriber, however, has no choice in what messages are received. In order to implement this, a broadcaster would have to expose multiple public feeds, which would be subscribed to by followers. An ‘aggregate feed’ could encapsulate all public feeds. The subscriber could retain the Circle-like privacy functionality by defining a subset of the user base to whom to whom to send a message. The subscribers which receive the message would be those who exist in the intersection of the targeted subset and the subset of users subscribed to the feeds the message was transmitted on. This model would retain the privacy control of Google Plus, while providing a stronger message filtration model than Twitter.

Comments

Leave a Reply

Blogging Calendar

September 2011
M T W T F S S
« Aug   Oct »
 1234
567891011
12131415161718
19202122232425
2627282930  

Archives