Some of my argument is indeed semantic. I don’t really know the underlying meaning of “date”, as I’m British and we don’t have quite the same culture, i.e. set of assumed meanings. If I start using the term “date” it will start to carry different meanings that a machine isn’t going to be sensitive to – that’s really my point. The point of XFN surely isn’t the act of marking up – it’s about the network you can infer. And the strings used as rel values I’d suggest are important – language is complex, and we use different words in our spoken or written for real reasons because the underlying meanings are never the same.

I understand that there was an element of Chris Messina’s argument that ran: “No-one is using the other rel values in large numbers. We should concentrate on evangelising the ones people are using because harping on about these other ones is just muddying the argument.” I was talking more about why people aren’t using the other ones. It’s because the unwashed masses can see them with clear eyes, and they just don’t make sense to them. I was trying to suggest why this is.

Put it this way: most people tag photos on Flickr
because they see it as a benefit, in general, that photos are tagged and therefore indexable by machines and searchable by humans. Why would people make aspects of their relationships explicit just so those relationships can be indexed by machines and other people can search them giving a set of search results devoid of the context in which the relationship statuses were declared? It’s necessary to explain to people, in terms not related to specific technologies, why doing so increases the common good in the same way tagging content does. And “because it’s XFN, because it’s a microformat” doesn’t count as a reason. Even most “technologists” on this globe have never heard of microformats.

So my argument was more qualitative than quantitative. I reject any idea that XFN presents a good that will be realised when more people out there are “educated” enough to understand the utility of marking up relationships explicitly.

I’m definitely interested in how you use the other rel values, and how they are useful to you! Do you find they help you to organise your own contacts? How much meaning do you think they provide for other people? Does the fact that the rel values are machine-readable help you at all?

I want to say I did a lot of thinking on these issues while writing this post, and only some of my thoughts found their way in. I should have mentioned my plan to set up my own crawler site that used XFN links and expressed the meanings they might infer, to the point of absurdity. “Alice is dating Bob but Bob isn’t reading so much into it and just wants someone to talk to.” Then I found RubHub. Well, it would have been a lot more exciting than RubHub.

As for suggesting something die without giving a possible replacement, I’m sorry I came across as negative – not my intention. At one level, my argument is that XFN is an unnecessary layer. The replacement is inferring relationships implicitly from what we are observed to do in different contexts. The replacement is affixing words to describe those relationships a posteriori. That’s where we need to go, and that is, I would suggest, what is currently evolving. Sticking to a flawed spec from five years back is holding down the debate.

Chris’ point was more that most people just care about ‘does know’ and ‘does not know’, thus the rest can be temporarily ignored until such time (if it comes) when most people care about more.

Of course, then there’s weirdos like me that care about more. contact, friend, kin, and colleague are all useful to me, and I use met, sibling, and coworker as well