I've been marking my internet territory. That is to say, I've been federating my identity. That is to say, I added more rel="me" links to the Larry Hosken contact page.
Do you remember my ramble about XFN a few weeks ago? Or the follow-up ramble about how XFN-like links might ease the burden of finding your friends on a social network service you just joined? Or how I set up that FOAF profile back in 2003 and then whined because no-one else had a FOAF profile so I had no-one to link to?
I hope you don't remember those. They were ranty and rambly. Which is too bad, because this stuff got more interesting recently.
Here's a summary.
Probably more than one web page represents you. Maybe you have a blog, a flickr page and a twitter page. Wouldn't it be nice to specify that those pages were all associated with the same person?
You can--if you can edit the page's source code. You can add a rel="me" to the link. If your blog links to your flickr with rel="me", you've "claimed" your flickr. If your flickr in turn links back to your blog with rel="me", that confirms the claim. If you can't edit the page's source code, then you might still get a rel="me" link, if the service lets you specify your web page and then automatically supplies that rel="me link. LinkedIn does, flickr does, other services do, too.
You can identify your friends. By adding rel="met", rel="met acquaintance", rel="met friend", or similar attributes to a link, you can say "The web page at the other end of this link is a person who I met/liked/whatever". (For a list of the words you can have in the rel, read the XFN spec).
A while back, I added rel="met" to various links on my friends-links page. (I didn't try to add further qualifiers like friend or acquaintance because... I dunno, it just seemed a little too high-school.) Social networks sites can use these, too. LiveJournal annotates friends links with rel="friend"... uhm, in some places. Not on my LJ page, apparently. But on some other people's LJ pages.
So, what's new? What got me excited about this stuff again?
This info just got easier to use. Google crawls the web. Google recently started keeping track of these XFN links (and FOAF profiles, but... this "summary" is already getting long.) Google is making this info available to programmers via an API and to humans via a sample program that uses that API. The API is Socialgraph.
- Socialgraph's list of my sites, with those sites it's most certain of listed first.
- Socialgraph's list of my friends: harder to wade through, since it doesn't filter out the stuff it's not certain of, but I see that Charles Ying has an XFN-enhanced linked to me--as do three Livejournallers.
Yeah, that information is easily accessible now. A programmer who wants to use it doesn't need to set up their own crawl. Yay, Socialgraph! I also like the fact that it uses data from more than one source: FOAF in addition to XFN. I'm not super-fond of either standard; it's good that Socialgraph is flexible enough to work with either; presumably it will work with other, better standards that come along. Oh, and the docs have pretty diagrams showing how a social network could use this info to solve the new-member-wants-to-find-already-subscribed-friends problem.
Full disclosure: Yeah, you know where I work. I hope that doesn't bias me in favor of Socialgraph, but you never know. My opinions are, as always, mine. I don't speak for anyone else.
Labels: site, social networks