Matt McAllister posted yesterday, challenging why (and how) people tag things. Muhammad Saleem posted yesterday on creating a meta-utopia in a controlled environment. Whitney McNamara posts today, to let you know that you should read both items above, and to note some of my thoughts that they inspired.
To summarize in a prefunctory fashion (which is why you should actually read both posts linked above), both Matt and Muhammad have concerns about the unstructured nature of arbitrary user tags as metadata, with varying ideas on how we might provide structure. (Matt looks in the direction of machines feeding tags to the user, Muhammad suggests an editorial layer of humans above the tagging users.)
Me? Let a thousand taxonomies bloom, man. I’m not particularly concerned about my tags being helpful to others, and I have even less interest in my tags being “right.” Tagging is useful to me precisely because it’s unstructured: if I had to remember some predefined system of classification that was comprehensive enough to cover all the Web content that I tag, I’d…well, I either wouldn’t do it or I’d tag everything with the small subset of classifications that I could easily keep in my head.
That situation would suck for me, because the tags wouldn’t be particularly good for helping me locate the items I’d tagged, and it would suck for everyone else because my tags would either be so generic as to be useless or so inaccurate as to be useless. A sweet little lose-lose scenario.
Now I’ll happily agree that tagging generates a lot of meta-noise along with the metadata. If you’ve got a small group of people classifying, you could be better off defining your classification structure up front: one of the odd things about tagging is that it seems to provide the best return at the very micro and very macro levels.
Picking the arbitrary tags that make sense to me works for me, because those tags fit the way I think. A million people picking arbitrary tags that work for them individually also works for me, because we start to see where all of those individual taxonomies intersect. Ten or a hundred people picking arbitrary tags…possibly a problem. Each individual still gets value out of their own tags, but the hive mind probably isn’t big enough for the commonalities — the common sense classifications, if you will — to start bubbling up to the top.
As with so many other things, I end up a centrist: as much as I enjoy debating the merits of X vs. Y, when it comes down to it I want both X and Y. There’s a reason that I own four different kinds of hammer…