On Apr 29, 2007, at 4:35 PM, Andrew Gray wrote:
And here we come to the fundamental problem. It's all very well to say we need to redevelop the category system, to say that the system is broken and failing and so on, but in order to do this we need to *decide what categories are*.
We have yet to manage this. Everyone brings their own conceptions to the table; we regularly get complaints about the stupidity and wastefulness of the system, which are fundamentally just "my preferred concept of what this system is is not being used".
We need to sort this out before we go any further - and the problem is that both sides are entirely convinced that theirs is the self-evident only way to work.
FWIW, I'm unconvinced we do that badly at this. (And I'm actually a recent convert - playing around with Catfishing has made me suspect that the categories are better organized than I'd previously assumed)
Let's look at [[Peter Pan]] (Randomly chosen). 8 categories.
Children's Novels, J.M. Barrie Plays, Peter Pan, British children's literature, fantasy novels - all obvious choices.
Literature protagonists is the only ridiculously broad one, being a category that should contain one article for every piece of narrative literature ever published. It's not useless as such, but it needs a refocus.
Literature featuring anthropomorphic characters, fictional characters who can fly, Kingdom Hearts characters - all three fall into a border zone - they're interesting to some people, could be used well by some people, but are probably a bit esoteric. Their flaw is not that they don't belong, but rather that they're on a threshhold of usefulness that if we took everything from that level of usefulness we'd flood the article.
There are a few categories I'd still love to kill - the birth/death years are the biggest. But most of them serve some purpose.
The problem, as I see it, is primarily implementation - the fact that categories are attributes of articles, and so every category must appear on the article. To my mind, categories are most preferable as replacements for the old "list of X" articles - navigational tools that work well on their own. But instead they're enormously difficult to change, requiring editing of hundreds of articles for a large category.
If categories could be moved to their own, independent existence and if a decent interface for the article side of things could be devised we'd have something that is very close to a useful system.
-Phil