I have a love/hate relationship with FOAF.

Yes, I’ve got my own FOAF file (permanently linked in the lower right of my blog gutter, in fact). Indeed the ability of people to own and maintain their own FOAF files, and for scutter/crawlers and find these files and make N-degrees-of-separation graphs and other interesting connections/deductions, is among the potentially powerful and endearing features of what I like to think of as ‘the FOAF promise’.

You can use FoaF Explorer to browse my FOAF neighbourhood. You can find a ‘FOAF codepiction photo path’ (a chain of photos known by means of some annotation to have the named individuals within them) that links me to Frank Sinatra.

Alas, there are problems.

• For starters, no less than 4 out of 4 of the first Google finds for foaf browser either have disappeared, don’t run at all, or fail to find people whom I know have FOAF files. Grrrrrrr.
• The FOAF files of people whom I know to be staunch FOAF evangelists are woefully inaccurate and out of date. Most are ‘trial runs’. My FOAF file lingers there, kind of like my old and no-longer-used Orkut subscription.
• People who inherited auto-generated FOAF files when Tribe.net became FOAF-compliant now have two FOAF files, often wildly different.
• I’m not even going to address the unagreed standards and the debates over the semantics of ‘knows’ and ‘has-met’ relations (these may ultimately be resolved, and more lightweight standards such as foafnet are likely to grow in popularit.

In contrast to the above promises-riddled-with-woes (hey, I’m a patient guy, but this saga has been going on for at least five years… and that is a looooooooooong saga on the Internet), consider the improbable case of un-owned and often un-attributed (provenance-free) personal data, editable by anybody/everybody, yet demonstrably stable, sound, and scaleable. Yes, my friends, we’re talking about Wikipedia, noteworthy for its use of swarm intelligence to maintain and update entries.

What has this got to do with FOAF and digital identity?

I notice on wikipedia that there are no less than 13 different people named Michael Jackson who have wikipedia entries!!! Yet the ambiguity is easily and nicely dealt with: if you go to the ‘main’ or ‘naive’ Michael Jackson entry the first line tells you “For other people with the same name, see Michael Jackson (disambiguation)” and you follow that link you get to a nice clean listing of the full contingent of 13 Michael Jackson entries, with their own unique URLs. In no case is there anything even remotely resembling FOAF’s popular ’sha1sum’ hash-coded email address which maps onto a provably-unique digital identity.

Who is ‘Michael Jackson’ then? Who owns these wikipedia entries?

Answer: No one / everyone, in true wikipedia style; and it seems to work remarkably well! There’s a fantastic lesson buried in here for FOAF/rdf/digital identity hackers, and it so simple as to be rather embarrassingly obvious: You are not, and your friendship relationships are not, nor will they ever be, your FOAF file, no matter how slick, detailed and up-to-date it is. You are probably not even, contrary to current wisdom, what you blog. More likely, in the best tradition of wikipedia and swarm intelligence, you are simply what other people think you are.

[UPDATE: I had intended to mention Stowe Boyd’s recent commentaries on unlinking from Social Networks, and his pointer to Mark Pincus’s thoughts on PeopleWeb, addressing many similar issues. The fantastic paradox is that the wikipedia-style described above simultaneously centralizes (via one common URL) and distributes (via the chaos of swarm intelligence authoring) the ownership.

Technorati Tags: foaf, digital.identity, swarm.intelligence, wikipedia

You can leave a response below, or trackback from your own site. You can follow responses to this entry through the Comments (RSS 2.0) Feed.