Posts from Friday, March 11th, 2005

Social Protocols

Published 19 years, 1 week past

Seems like half the Web is already at SXSW, and I’ll be there myself soon.  For those of you who love to build networks out of your social contacts at such events, Tantek’s recently shared the secret of metrolling, which is a great way to get into XFN if you haven’t already.  I’m already planning to add metrolling to my presentation on Sunday as an example of ground-up semantics.  (And I really wish I could be at the Semantic Web panel on Monday, but it’s at the same time as “Women of Web Design”… oh well.)

It’s interesting to see how interest in evolutionary semantics is itself evolving.  A recent example of this is David Berlind‘s ZDNet article “Will social networks give way to social protocols?“.  I firmly believe the answer to be “yes”, even though there are a lot of skeptics (some of them on conference review committees, as it turns out).  Berlind clearly understands the advantage of social protocols.

You might then wonder, “Then what’s up with you writing a whole document about how to set up XFN ‘me’ values in a bunch of services?”  At this stage of social networking, that sort of thing is necessary.  Without interim steps, the information sitting in those services will stay scattered and isolated.  Thanks to the me value, XFN offers a very simple, lightweight solution to the problem of identity consolidation.  As I recently wrote in a poster proposal:

As the Web has evolved, a number of personal-information sites have arisen.  Some of these sites exist to help create and increase professional contacts; others are intended to help bring together one’s friends or even find potential mates.  In every case, however, the user must create a new profile for each site.  Each of these profiles constitutes a small island of identity.  Over time, a person can end up with a fairly extensive identity archipelago.

Unfortunately, there has… been no easily created machine-discoverable way to bridge the gaps between the islands.  An author might publish a page containing links to all his profiles, but to an indexing engine, these links are no more or less notable than his links to the latest amusing Flash animations.

With XFN, it becomes very easy for an author to annotate a link to indicate that its destination is one of the islands in his identity archipelago.  This kind of link is referred to as a “me” link throughout the rest of this paper.  By creating symmetric links between the islands, the author can make it possible to consolidate the various pieces of his online identity into a more cohesive whole.

The same is true for a person’s links to other people.  By pulling them all into one place, or at least by marking them all with XFN and then using “me” links to tie together all the bits of his identity archipelago, real social networking start to emerge.

Now, one of the things that people like to carp about is the limits of XFN.  The first of the two most common complaints are that it’s impossible to capture the full range of human relationships in fifteen words.  We agree.  The other complaint is that we only picked “positive” terms; that is, we have friend but not enemy.  We did that on purpose, as we explained; besides, it’s called XHTML Friends Network, which should be kind of a clue.  Apparently this choice makes us arrogant, or clueless, or some combination thereof.  Maybe that’s so.  What I find interesting is that the people who complain that we didn’t include their preferred relationship terms never do anything about it.  They just complain.  What’s so interesting to me is that the guys who decided to focus on the positive went out and did something; those who want to mix in the negative seem to have nothing to offer except complaints.  That says something, I think.

Because XFN is not, nor was it ever meant or represented to be, the final word on social protocols.  We fully expect it to either be improved, or else superseded.  Suppose one of the critics actually did something to address his concerns, and published an “XHTML Relationships Network”.  This could include all the XFN values, plus their negative counterparts, plus whatever else is thought to be useful by the author(s) of this new XRN.  At that point, you have competing protocols.  The more useful one will win.  The loser will be eventually discarded, although some of its memetic genes may live on.  This isn’t a problem: it’s a strength.

It’s also in many ways the entire point of XHTML Meta Data Profiles.  See a need to fill?  Fill it!  At the end of his column, Berlind says in an update:

Looking at the XFN profile, it suddenly dawned on me that perhaps there should be an XBN/XB2BN that’s strictly for the relationships between businesspeople/businesses. Thoughts?

Here are my thoughts: go for it!  He’s almost certainly right that there’s utility in such a protocol.  All it takes now is for someone to look at the problem and write up an XMDP-based protocol that solves the problem.  The microformat approach makes this so simple, pretty much anyone could do it.  What’s needed is someone who actually will do it.

At some point down the road, it’s possible that the protocols that define personal and professional relationships would merge.  Again, that’s completely in keeping with the vision we have.  The whole point of this kind of ad-hoc semantic enrichment is that it’s evolutionary.  New players will enter the field, and will either prosper or wither.  Anyone can join in.  There is no star chamber of lofty experts to say whether your idea passes some sort of ideological muster.  It’s a great big landscape, and there a million conceptual niches to be filled.

As those niches are filled, the ways in which different protocols interact can trigger truly astounding results… but for thoughts on that aspect of the whole subject, you’ll have to come see my talk.


Browse the Archive