Mining passive social networks

Recovered from the Wayback Machine.

Marc Canter pointed to an upcoming conference, the 1st Workshop on Friend of a Friend, Social Networking, and the Semantic Web (also mentioned at RDFWeb and Danny Ayers). I wasn’t sure about a conference focusing primarily on Friend of a Friend, but looking through the Call for Papers topic list, the conference goes beyond the concept of friend of a friend, if we think of this as primarily focusing on people describing their current, active associations. It also has the potential to go beyond, I think, the current understanding of social networks.

Though social networks have past the hot, new meme phase, we’re still trying to figure out everyday uses for these beasties. Or I should say, I’m trying to find practical and unique uses for social networks. If they continue on as they are, they’re really nothing more than a combination of popularity contest, dating service, job search, mini-blog, and Usenet.

However, the thought of associations defined online in a format that can be mined using standard technologies has a great deal of appeal–and not just within the context of making associations for jobs or professional advancement. If we look at FOAF as a way of enabling open source social networks, and then extend the definition of social networks to include passive associations rather than, or in addition to, active associations–then we have something really unique, and rather exciting.

For instance, I’ve been trying for the last year to find out what happened to others who graduated in 1986 with computer science degrees from Central Washington University. The Alumni directory of CWU isn’t particularly useful because it mainly lists information about people who actively post some new events about themselves, and most of us don’t. I also don’t have access to classmates’ addresses (the Alumni association is wary of giving these out); even if I did, I would hesistate to send them notes because I wouldn’t be sure if these people would want the contact with their old classmates.

If I modify my FOAF file to include information about me attending Central Washingtion University and getting a degree in Computer Science in 1986, this is a node, albeit a passive one, in an existing social network comprised of people who I went to school with. Now if this information were specifically searchable, someone else wanting to connect up with fellow classmates would find my information, including my email address, and they could send me an email, saying hi.

How is this different than the means I have today to connect up with others of my old class? It differs in two important ways:

First of all, the information is publicly available and searchable on the Net. Not only that, but it’s also constrained to specific assertions: I have a degree in computer science; this degree is from Central Washington University; I received this degree in 1986. Someone searching on just these assertions would find my data, rather than get 10,000 records back with a lot of accidental associations based on random scraping of pages in Google.

Secondly, and most importantly, by putting this information into my FOAF file I am acknowledging that I am a part of this particular social network, with an implication that I would welcome associations from other members of this same network. It is a passive node in the network, true; but with FOAF/RDF/OWL and data mining, these passive associations can be converted to active ones in the future.

Orkut leaves me lukewarm and Friendster leaves me cold. Even the Semantic Web doesn’t do much more than elicit a relatively unenthusiastic “yeah, smart shit” out of me. But give me this functionality in the future and I’ll dance from dawn to dusk.

Hopefully Workshops such as this one in Ireland (oh, I’d love to go), will help make this functionality a reality.

Print Friendly, PDF & Email