XFN Foofiness

Let me state that this just occurred to me, so it’s going to seem a bit weird.

In making a comment on Jeff‘s site, I almost did my standard markup that occurs to me when I type Amy‘s name, namely:

<a href="http://domesticat.net/" rel="friend met">Amy</a>

Now, here’s the thing … Amy is my friend, but she’s Jeff’s wife. If I comment that I make on Jeff’s site, is some theoretical Semantic Web content-extractor going to think that Amy is Jeff’s friend whom he’s met, or will the rel=”friend met” apply to me, since I am the author of my comment?

This is not a very trivial matter. If Jeff got with the XFN and started putting rel=”friend met co-resident spouse sweetheart” into every link with Amy’s name in it—admittedly, a bit of a chore unless Jeff hacks together a glossary of people that he mentions semi-frequently on his site so that a little hackery allows WordPress to auto-generate such gloriously laborious information—will a SemWeb c-e get confused by this?

Has this occurred to anyone else? :shrug: I just post crap as I think about it.


  1. I think Matt works on the XFN stuff, you should put this on his radar. I think it’s definitely an interesting scenario – the comments and posts need to be somehow related to their respective authors I guess.

  2. After I posted it, I showed it to Matt. He noted that folks really should only XFN on their own sites. I get his point, but to be honest, I consider the comments that I make on others’ sites to be mine and not theirs—it’s something that I’ve contributed, ya know?

    It’s just one of those trial balloons that popped into my head. It’ll be interesting to see how much XFN takes hold in the next few months, eh?

  3. Currently the rubhub.com spider will assume that the rel attributes are authored by Jeff (because they’re on his site) and will add any new ones that you’ve put in there. You’re right though, it could be a big problem. Unfortunately, differentiating comments from posts in a page of html is a whole other level of complexity away from what I’m planning for the next iteration of the spider. Thanks for putting it on my radar though. As a site author, I think I’d want to be able to automatically strip any rel attributes out of links in comments in order to avoid that kind of issue.

  4. Phil:

    Thanks for your comments. Happy to help.

    What makes me wonder here is if there’s a way to generally differentiate comments from post text. I can see some possibilities offhand, but it sounds to me like a regexp nightmare. 🙁

Comments are closed.