I don't like bandwagons, because I am a slowcoach and it's usually pretty crowded by the time I get there. However, reading through the info on XFN, I don't think I'm going to be able to keep my trap shut on a couple of issues.
At first glance, we see that it is an "elegant" solution: it's simple, it's readable, and it's XHTML. However, let's say, out of the context of one's blogroll (a word whose definition is outside the scope of XHTML, and thus outside our boundaries of enforcibility), there's no restriction on using XFN. What's wrong with that? Well, ever wonder about the person whom you thought you were close to, but haven't spoken for three years? Is that person still a friend? Probably less of a friend than a colleague you get along well with and speak to every day. But three years ago, you had all these XHTML references to a person with whom you are no longer a friend. I don't know about you, but I see a major maintenance issue, on top of trying to keep up with all my "friendships". :)
The authors of XFN have eliminated the "time" factor in the interest of "simplicity". Simplicity, for the sake of accuracy? I find this somewhat ironic, given that XFN attempts to strive for a kind of clarity in describing relationships. Is there use in describing relationships that are no longer accurate?
Truthfully, this may not be unfixable. It doesn't take a database or IA expert to point out that this is where one would use a unique key to correspond to a person. Instead of describing the relationship in the anchor tag, say, we use a unique id; of course, if we do this, the amount of of parsing involved in order to build a relationship graph is at least twice as complex. In any case, isn't this what RDF is for? So why are we using XHTML for something that it is not made for? (Why hammer a nail with a pair of chopsticks?)
Technical problems aside, I see internationalisation issues in enforcing gender neutrality. I shall provide two examples. First: the Chinese have very clear descriptions of family relationships. I have a precise term for the the second oldest uncle on my mother's side of the family — it is 二舅. I would correctly call a female cousin who's the child of my uncle on my father's side who is younger than me: 堂妹, but a female cousin older than me who's a child of my aunt on my father's side is correctly called: 表姐. As you see, in this context, precise relationships are important. Once you get down to specifying a relationship beyond "knowing", therein lies a quagmire.
The second example I have is somewhat reversed to the first. In a good part of South East Asia, it is customary to call someone on the street "aunt", "uncle", "grandfather" or "grandmother" even if they are not related to you by blood, whether out of respect or affection. I have known "Auntie Monica" all my life, but when I meet a Westerner I am forced to explain that she is not related to me by blood, but she is an aunt to me in all other respects — what I mean as an "aunt" differs considerably in meaning to that of someone else; however, my aunt here is just as important to me as my parents. I can't describe this relationship within XFN? That's a shame.
The XFN specs have repeatedly said that one defines relationship according to the author's perspective. So someone I claim as my friend may not necessarily think of me as a friend. Is there truly value in this? Given that temporal factors are not considered — isn't it true that I will always be accumulating friends! Yes, okay, that's not a bad thing for my ego — but how meaningful is this information?
Needless to say, I also have some philosophical issues with the approach XFN has taken. Aside from the contradiction in striving for accuracy (in defining relationships) yet never actually being able to be usefully accurate (due to temporal and cultural limitations), I simply have difficulty to understand that there is any practicality in categorising my relationships according to an arbitrary set of terms.
Complex problems don't often come with simple answers. Relationships are ambiguous, complex, they change, and sometimes escape definition. Why oversimplify them to such an extent?
Posted by sniffles at December 17, 2003 10:34 AM