Is XFN worse than FOAF?

You might think that XFN would be even more even more decentralized and emergent than FOAF.
Both are decentralized ways for individuals to describe their social relationships, in contrast to the centralized social networking services from Friendster et al. XFN uses hyperlinks to describe the linker’s relationships with the linkee, while FOAF (friend of a friend) uses a file in XML/RDF format listing all the friend relationships.
The reason that hyperlinks are generally such a nice way of showing emergent patterns is that they reflect millions of tiny choices that individuals make about what’s relevant.
But the XFN site anticipates that the primary use of XFN will be in blogrolls. If that’s the case then the relationship is buried in markup, in a list that doesn’t change very often. This replicates one of the major problems with relationship profiles — they are static, while relationships change slightly with every interaction.
Am I really going to update my blogroll to add a “met” attribute for Kevin Marks after meeting him in person at Etech? That kind of micro-maintenance will happen even more rarely than people clean their closets.
Kevin Marks’ “vote links” are a much simpler and likelier use of expressive hyperlinks to show emergent opinion. The choices are simpler — vote-for and vote-against. They let users express distinctions they want to express. In an article about Diebold’s ghastly security holes, you can link to Diebold with a vote-against link.
By contrast, do people really want to declare a relationship as a “crush”, “date”, or “sweetheart”? (These are real XFN vocabulary terms). Imagine the agony of deciding when to switch from “date” to “sweetheart”. Is it after the first kiss, or the third date, or flowers, or what?
Link emergence works when the links are frequent and simple. XFN won’t create link emergence, because the links are static and complicated.

3 thoughts on “Is XFN worse than FOAF?”

  1. everything changes… the problem with blogrolls is that they start out with best intentions but often fall into a static state of meaninglessness… how many people’s blogrolls bear even a close resemblance to their newsreaders? now a system that would automatically ping me as it sensed that my links were going stale based on preset variables of time, links, comments, trackbacks, etc., and auto-encode my responses in markup would be most valuable – help me garden out abandoned weblogs, changed urls, changed interests from my links, blogrolls, newsreaders,… btw, i vote for the voting concept – for, neutral, against… (-:=

  2. I’m starting to think that FOAF systems (generically speaking) are a solution in search of a problem. But an additional problem with FOAF (specifically speaking), OPML and apparently XFN is that a nice standard is only one piece of what’s necessary to get a decentralized system of any kind off the ground. You also need tools.
    As I said at the Austin bloggers anti-Meetup the other night, a logical place to turn OPML or XFN or some other standard that can be used for blogrolls into something useful is as a plugin to Moveable Type. Moveable Type has a broad enough installation base and blogroll sidebars are a desireable enough blog feature to motivate people to load up their (blogosphere-centric) contact info into it.
    And I argue that you *might* weed your garden and harvest the blooms a little more often if it were a matter of pointing and clicking, like the silly ice cubes/smilies/hearts interface in Orkut.
    (There remains the question, though, as to why anybody would want to syndicate anyone else’s blogroll. Other than the Department of Homeland Security, maybe.)

Leave a Reply

Your email address will not be published. Required fields are marked *