Community and social computing
with Tags: semantic_web X
Reading this story on Red Herring about the Socialstream project over at Carnegie Mellon, reminded me of three other things. The story is about a graduate student project that aggregates "any" social networks. What that means is obviously vague but it's getting some press due to the Google funding it received. It partially hints at the ability to aggregate a user's profile from across multiple sites. That in itself is not that unique.
The first thing it reminded me of is ZoomInfo, a site that conducts a web search for a person's name and makes correlations, and creates a single profile for all the related info that match that person. It isn't all that accurate and sometimes gives multiple hits for the same person (e.g., "Rawn Shah" a name that as far as I have searched is currently unique in the world). It also can give results for multiple people (e.g., plenty of "Michael O'Connell"s out there), and I'm guessing that sometimes the entries cross over between profiles inaccurately. It doesn't quite merge profiles from multiple social networks as much as take raw information, do some semantic processing looking for similar context, contact information, and references.
What would help ZoomInfo is more contextual or semantic declaration of information, in other words, it is an application just waiting for the Semantic Web to rise. There are other examples like the real-estate mapping and pricing site, Zillow. Both search multiple sources for particular type of information and then aggregate them in some meaningul way to do what they do. I know what you're think: "Isn't that just a mashup?" And you'd be right, but this kind of mashup becomes more useful if given the semantics of the situation.
This leads me to my second thing, the Semantic Web, which we talked about just recently at work. Conceptually, it means applying more specific information to any piece of knowledge so that it can be handled by software agents trying to understand the context and semantics. This is the current work of Sir Tim Berners-Lee, the originator of the Web. I think we even have a podcast interview him on dW where he discusses this. This has also been referred to more recently as "Web 3.0", although the Semantic Web seems a more accurate and less self-aggrandizing title. To get to this level, there are a number of protocols, in particular the Resource Description Framework, that need to be much more widely applied to web information. There's a more complete diagram on that Wikipedia entry for it.
But returning to the topic of "aggregating" social networks, the third thing I recall is what Marc Canter, founder of MacroMedia and current CEO of Broadband Mechanics, wrote on CNet: "...it's clear that folks are interested in connecting together some of their disparate accounts on a wide range of social networks." Having worked on standards before I know how much messy politics go into it more often than good information. Marc's focus here again is also on connecting the profiles across multiple networks, and separately connecting the social networking tools they use and the information they generate.
My point of view, all three (profile, SN tool, information) are part of the same set, and should probably go together, not in terms of actually being on the same application, but in terms of the same presentation. In the background, all the data in the set can come from multiple application sources, even multiple sites. Don't force people to switch tools and interfaces constantly so they have to go from e.g., their blog to their wiki to enter different kinds of information. The same for profiles, you can have multiple profiles on multiple sites, but even switching between different sites to get to your profile in each community/social network system you are part of is cumbersome, when you need to change some basic information, e.g., your job title. It's also hard to track.
Is there a solution?