ISS was designed with one interesting characteristic in mind: incoming information only comes from one-degree apart friends through subscriptions. This characteristic provides an architecture which is 100% guaranteed to be SPAM free. You’ll only receive information you want. And if you start receiving information you don’t want, you can either directly contact your friend and talk to him/her about it, or simply unsubscribe from him/her. It’s interesting to observe that the cascading of trustful social networks works as a world wide distributed recommender system perfectly tuned to output a very personalized stream. This trustful network filters out irrelevant information, while still letting good information pass through. Your friends work for you, and their friends work for them, and so forth. Everyone works together to recommend and filter information. ISS effectively unleashes the great deal of untapped potential of the collective intelligence to organize information.
The guys from the Decentralized Information Group at MIT have been pondering about this idea of combating SPAM using social networks. They implemented a FOAF crawler that generates a social graph periodically. And from this social graph they create policies around it, such as, for example, only people x-degrees apart can comment on the blog.
This goes very much aligned with how ISS was designed. When a user receives an entry from a friend, she can read the comments that her personal network has made about it. She may also download the attached files directly from her personal network, making it easier on the original author’s server in case the entry gets popular. And since the information that reaches the user always leaves a trace, it’s very well possible to choose, for example, to read only the comments for a particular entry from people x-degrees apart.
I updated the ISS policies to make it clear how ISS controls the quality of information that reaches the user. I’ve added a special case, which are friendship requests, since these requests obviously will come from people who are more than 1-degree apart. The funny thing is that eventually we’ll only add friends that are already on our list (not our personal network obviously, but our more wide social graph that was generated beforehand). Friendship requests from people more than x-degrees apart will be dropped, not because of despise, but simply because of trust. The requests that pass through this policy will be displayed to the user as a special channel, where the user can easily see the person’s tagcloud, how many degrees separate them and how much they are connected. He may add this person to his personal social network if he sees it appropriate.
These policies will initially be applied to the Web and I.M., but they’ll be applied to e-mail as well. These three means of communication will eventually converge and have a common interface. We will finally overcome SPAM and information overload, and reach an equilibrium of information-sharing and awareness.
