Human-readable data serialization language for Perl, Python, Ruby, Java, etc.

Ontology is Overrated: Links, Tags, and Post-hoc Metadata "The problem ontology solves is not how to organize ideas but how to organize things -- the Library of Congress's classification scheme exists not because concepts require consistent hierarchical placement, but because books do."

/me pontificating on the question being begged Turns out that the linked page is not quite as stupid as the headline makes it seem. But nevertheless, I think it would be good to realize that this is all a question of understanding your audience's needs. Audience: You yourself, or a close-knit group of people. Needs: None you can't handle yourself anyhow. Strategy: Don't bother to tag or organize anything just for the sake of being organized. Audience: Any size. Needs: Data is easy to classify unambiguously. Strategy: Just do that then. Audience: You yourself, or a close-knit group of people. Needs: A significant amount of data but not important or complex enough to spend "real effort" on. Strategy: Come up with some ad-hoc tags, and try to stick to them. Audience: A larger or more diverse group. Needs: Data is hard to classify unambiguously. Strategy: This is not a solved problem, but ontology, controlled language, terminology, taxonomy, etc are some of the strategies we have. It's never wrong to spend more time on using a more advanced strategy than you absolutely need, if somehow you can find the resources to keep on doing that. Needs evolve and it's very hard to upgrade to a better strategy later, so if you can imagine that what you have now might not be sufficient one day, it's a good idea to spend some extra effort now. Who should be responsible for standardizing tags, when that need exists? I don't think that's an important question. When a tag is already in widespread use, and/or likely to be the one your audience will search for, it would be stupid not to use it, though. This can incidentally be generalized to another principle -- if some other categorization system already exists for the data you are tagging, it's likely that you are doing something wrong. (But of course, using tags in parallel with some other system might make sense.) The other mistake to watch out for is whimsical tagging. It scares me to see what tags people use (that includes my own tags, when I revisit them, more often than I like). If you don't know what a tag means, don't use it, or find out more about it.

Ha ha, again somebody explaining why Clay Shirky is wrong

An introduction to DOAP with a working example