Developed with personal weblogs in mind, XHTML Friends Network is a way of identifying human relationships within a Web page's code. When you link to a friend's Web site, throw in a
rel="friend" to declare your relationship to the world. That'll allow you to target those links in your own code (perhaps to offset them in some way), and it'll provide all the necessary information for yet-to-be-written information-gathering robots to...well...do something.
What, exactly, can be done with this data is anybody's guess. Two ideas I've seen so far: "Trust networks" and dynamically rendered asterisks beside links to real-life friends' pages. The former is slightly pie-in-the-sky; the latter is...well...something.
Although I'm hard-pressed to think of an online-news application for this technology (21st-Century news media don't have personal relationships!), let alone any practical application, I've added XFN code to the blogroll on this site's home page. I'm a sucker for early markup adaption. And while I may have my doubts about the tangible effects XFN will have on the Web, it's clear to me that XFN is a good thing.
The significance of XFN is that it will get more people thinking about, and appreciating, metadata. The more people appreciate metadata, the less expensive it becomes to create. The less expensive metadata is to create, the more often metadata is used. And the more often metadata is used, the more applications will be invented to utilize it.
And, finally, that'll be something.