The Google Social Graph API: the good and the bad

I was very interested to hear about the launch of Google’s social graph API at the weekend. The social graph API “returns web addresses of public pages and publicly-declared connections between them”, where the connections are currently being obtained from crawled XFN and FOAF links. Dan Brickley, the co-creator of FOAF said:

The Google API looks like a step in a very interesting direction. Of course it will be possible to think of many things it doesn’t yet do, but I encourage everyone here to have a think about simple, practical and useful incremental improvements to it. We can do a lot more eg. with full SPARQL access, but proving full SPARQL to the aggregation of the planet’s public FOAF/XFN data isn’t going to happen anytime soon. Interesting times 🙂

In answer to Niall Larkin‘s question about how this relates to SIOC, such services help us because by providing an easy method to find one’s social graph (both “me” and “knows” connections), it also makes it easier to find your social objects which can be described using SIOC (see my previous illustration, and see also Kingsley Idehen’s demonstration of how this can work).

In short, you can use FOAF to create the social graph, and use SIOC to represent social objects.

Not everybody is entirely happy (see the comments on Tim O’Reilly’s blog post), with the majority of objections being in relation to the APIs being operated by a for-profit as opposed to a non-profit organisation, and there is some opposition to the idea of a single point of control rather than having a set of distributed indexes.

Perhaps we need something similar to “nofollow” links for the public social graph as well. We will discuss these issues and some other important social network portability topics at WebCamp SNP in four weeks time.

Edit: What I can add to this is my gut feeling that it probably requires a company like Google to make an API that can gather the required momentum and that people will use; previous FOAF aggregator efforts like Plink and FOAFSpace could have done this, but they would have found it much harder to gain critical mass.

Advertisement

Opening up the social graph at the WebCamp workshop on "social network portability"

A WebCamp “Social Network Portability” workshop has been announced to be co-located with BlogTalk on 2nd March 2008. You can view the wiki page for this event.

“Social network portability” is a term that has been used to describe the ability to reuse one’s own profile and contacts across various social networking sites and social media applications. At this workshop, presentations will be combined with breakout sessions to discuss all aspects of portability for social networking sites (including accounts, friends, activities / content, and applications).

Topics of relevance include, but are not limited to, social network centralisation versus decentralisation, OpenSocial, microformats including XHTML Friends Network (XFN) and hCard, authentication and authorisation, OpenID single sign-on, Bloom filters, categorising friends and personas, FOAF, ownership of your published content, SIOC, the OpenFriend format, the Social Network Aggregation Protocol (SNAP), aggregation and privacy, permissions and context, and the Extensible Messaging and Presence Protocol (XMPP).

You can register for this workshop in conjunction with BlogTalk 2008. If you are interested in speaking or otherwise participating in the workshop, please add your name under the Speakers or Participants headings on the wiki page at http://webcamp.org/SocialNetworkPortability.