The second age of the World Wide Web has brought a new social-media-centric vision of the world where users share information, interact mutually, collaborate. The hosted services that embrace Web 2.0 share more and more attributes: from early blogs to folksonomies, wikis, video sharings, mashups and, of course, social networkings.
Such a user-centered design along with the interoperation among users guide to the evolution of online communities, that spring up for every imaginable purpose, at any time, unlike in the offline world when it was more complicated for people to come together due to restrictions of place, time and money. And this web-based trend is just destined to increase with the advent of the so called pervasive computing, when the networking logic will absorb, through billions of ubiquitous robots spread worldwide, every frame of daily activity, in every location and in every context.
Given that there has been huge interest in using Web 2.0 technologies, I think we’re still stumbling around in the dark about how to best incorporate these new media strategies as part of a whole effective marketing strategy, how much time to invest on social networking, how to monetize the social media mechanism, etc. Even among those who are the most skilled in social media there are different approaches: there is who, like Chris Anderson (Editor-in-Chief of Wired magazine), sees social networking as a feature of a website, not a destination in and of itself. And there is who, like Charlene Li (Altimeter Group’s Founder and better known, here at the Future Centre, as the co-author of Groundswell), believes that social networking is becoming so important that it is like air. I like this similarity, at the point that I’m convinced that if we could analyze chemically SNs, we’d find all components of air pollution!
Joking apart, studying user interactions is a crucial task, and many efforts are underway to analyze them by analyzing SNs among users. For instance, it has been reported that an individual with friends in a group is significantly more likely to join the group if these friends are themselves mutual friends, rather than they are not. But despite what has been done, we surely need to keep experimenting and asking questions. What is clear, is that a social network has precise structural and organizational properties – for instance, network size and density, that respectively describe the number of peers and the general level of linkage among them – and that there are algorithms able to identify, starting from a given social network, important network-based features in order to analyze user behavior efficiently and to expand the services.
So, studying SNs becomes a way to improve effectiveness of communication infrastructures and, of course, the other way around, studying communication networks can reveal some interesting organizational features of the social-interaction processes, useful for predicting user interactions. Alessandro Panconesi, full professor at Sapienza University of Rome, is convinced of the connection between infrastructures and behaviors: in his research about the social dimension of the networking he explains the possible evolution of this phenomenon and the ideas that underlie it, so convincingly to obtain the IBM Faculty Award. That such a connection exists is quite self-evident: instant messaging and chatting mirror the immediacy of communication, SNs reflect the structure of the primacy we attribute to our own friends. But going into more depth, we could point out both interesting mathematical and social models!
One of the aims of this research, as a matter of fact, is to understand if, and the measure of that, the structural features of the SNs facilitate or slow down the diffusion of information due to “gossiping” processes. The term gossiping, borrowed by the social sciences, can be identified with the spreading of rumors in a network and refers to any communication algorithm where messages between nodes are exchanged opportunistically, with other nodes that act as betweeners or forwarders of the messages. The rumor spreading starts when the initiator node chooses randomly a friend to whom communicate the gossip; from now on, every peer that is acquainted with the gossip will do continuously the same thing, communicating the information to a randomly chosen neighbor, that will do the same with another peer, and so on.
A gossip-based algorithm is similar to that one inspired from the virus spreading, with the main difference that gossiping nodes adopt a human-like communication paradigm, while the infected nodes behave like pathogenic microorganisms (bacteria, parasites, viruses exactly).
The preliminary results of this research have already obtained a prestigious international award. The work “Almost Tight Bounds for Rumour Spreading with Conductance”, which Panconesi co-developed with his PhD students, has been invited to the 42th ACM Symposium on Theory of Computing (STOC 2010), the most prestigious international congress in the field of theoretical information science, that will be held in Cambridge, MA, from June 6th to June 8th.