Last week Flock sponsored the W3C Workshop on the Future of Social Networking in Barcelona, and I had the honour of attending on behalf of the company. The purpose of the 2-day workshop was to help determine what role the W3C should play, if any, in the emerging field of social networking services. The event was chaired by Dominique Hazael-Massieux and Christine Perey.
About 80 people were in attendance, and 72 position papers were submitted beforehand to help provide some background and context for discussion. That’s a lot of papers to read. I got through most of them, but not all, prior to the workshop. A nice short synopsis of each paper can be found here (and part two). There was also some ongoing Twitter conversation during the workshop and the #w3csn topic was trending on Twitter for a while.
The breakdown of attendees was roughly one-third academics, one-third mobile industry, and one-third “other” including businesses involved to various degrees in social networking. Conspicuously absent were representatives from any of the major social network operators; no one from Facebook, MySpace, Twitter, Flickr, etc. (YouTube and Bebo submitted position papers, but did not give presentations and I didn’t notice them in attendance. If someone from YouTube or Bebo was actually there, sorry I missed you!) Geographic representation at the workshop was also predominantly European.
Day 1 – Thursday, January 15th
Presentations and discussion centered around “architectures for social networking”. Many attendees (and indeed many of the submitted position papers) lamented the current “walled garden” model wherein social network operators are incentivised to closely guard their users, content and media. Rather than moving towards an SNS monopoly or oligopoly, the SNS landscape is likely to become increasingly fragmented. Despite the meteoric rise in usage of Facebook and MySpace in the last couple of years, other SNS operators are also seeing considerable growth. In fact, someone noted that the “long tail” of regional/corporate/special interest social networks accounts for about 500 million users! (Unfortunately I didn’t catch who said that or where the statistic came from.) Ultimately, users would be better served by a model that allowed them to have data portability between networks, to manage fragmented identities (by either combining them or keeping them completely disconnected, as desired), and to use different providers for the services and specialties they provide (eg. LinkedIn for resumes, MySpace for music, etc…) without being tied to particular network operators.
There were two concurrent breakout sessions in the morning: one on Distributed Social Networking (which I attended) and the other on Data Mining.
The Distributed Social Networking session had some high quality discussion, but did not result in any concrete recommendations as to potential roles for the W3C. Some architectures for distributed/decentralized social networking systems were discussed, and it is clear that the barrier to implementation is not a technical one. Existing data format standards, protocols and APIs such as OpenID, OAuth, OpenSocial, FOAF and XMPP are sufficient to implement such systems, but they don’t address the business forces shaping the walled-garden problem that is the status quo. That will require business model innovation, rather than technical, and as such isn’t really in the W3C’s realm.
The Data Mining breakout session apparently had more concrete results. The recommendation was for W3C to define a standardized data interchange format as a harmonization/extension of existing formats such as FOAF, Atom, etc. There were strong assertions from the academic crowd that RDF is the appropriate way to model social network data, but the idea of using it as an interchange format was generally poopooed due to its complexity.
After lunch, there were concurrent breakout sessions on Privacy and Trust (which I attended), and Distributed Architecture Business Models.
The Privacy and Trust session did not result in any strong recommendations for the W3C’s course of action. The issue of identity fragmentation (ie., users having multiple profiles on multiple services) was discussed at length. Some users consider it an inconvenience to manage multiple online identities, whereas others absolutely rely on it to maintain privacy — they don’t ever want their LinkedIn account to be associated with their MySpace account, for example.
The W3C’s existing P3P initiative was mentioned as needing to be extended in order to really be applicable for SNS sites. Other than that, user education/awareness was cited as the main issue needing to be addressed. One person said something to the effect of, “Wherever I go, it should be obvious to me what my current privacy and security context is.”
Blaine Cook (of BT, formerly of Twitter) threw out an interesting idea about using capabilities-based cloud data stores so that users could maintain better access controls over their data, AND the data could not trivially be associated with the user unless they wanted it to be. Worth some more thought.
The Distributed Architectures and Business Models breakout considered whether there are viable business models for a more distributed/decentralized model of online communities; in other words, can we escape the walled garden paradigm? There was apparently some disagreement between web and mobile operators in these discussions. Everyone agreed that selling in a social context has been shown to work. A sizeable contingent further believe that a widely adopted micropayments system would go a long way towards enabling economies between SNS operators, rather than just within their closed communities. In response to this, W3C may look at restarting its work on micropayment systems, which has been stagnant for a few years now.
Day 2 – Friday, January 16th
The presentations and most of the discussion on the second day of the workshop centered around the topic of “context and communities”. That is, enriching social applications and online social interactions with contextual data such as location, relationship, engagement mode, etc. There was a fair bit of discussion about geolocation and privacy controls for geodata.
Julian Pye of Vodafone gave an interesting presentation (paper, slides) on adapting user interface according to context. For instance, automatically stemming the flow of friend activity updates according to my relationship with the contact, whether I’m at home, work, or school, on a holiday or a business trip, my proximity to the source, etc.
Simon Hay from the University of Cambridge got more than a few chuckles for his entertaining use of Harry Potter analogies in his presentation (paper, slides) on the use of sensor arrays tied to social networks. They implemented something akin to the Marauder’s Map, which shows the location of every student on campus (or in the case of UC, just those students who chose to participate) and could notify you if, for example, several of your friends were heading down to the coffee shop.
There was also an interesting side-discussion about the validity of “Dunbar’s number” (usually cited as being 150) as the theoretical cognitive limit on the number of friends one can maintain. Harry Halpin asserted that this number was fallaciously extrapolated from studies on primate sociology, and that it’s much more useful to think in terms of a graduated scale of 12 intimate friends, 150 frequent contacts, 1500 infrequent contacts and 1,500,00 lifetime contacts as averages for humans.
As mentioned above, suggestions for the W3C’s course of action include:
- Put forward a recommendation for a data interchange format for social/identity data (as a harmonization/extension of existing data formats such as FOAF, etc.)
- Look at extending P3P to be more applicable to SNS sites and communities.
- Look at restarting Micropayments standardization work — there may be more business interest in this now than there was a few years ago.
In addition, Harry Halpin put forward a draft charter for a W3C Social Web Incubator Group to continue discussing standardization of the social web technology stack.
There was also a suggestion that a follow-up workshop be planned in 6 months’ time, probably in North America.
Update 2009-02-03: The official W3C report is now available on their site, and includes a couple of photos by yours truly. :)