Metadata, schmetadata … or useful context?
April 12, 2007 – 2:31 pmerp4it: A metadata rant is an example of folks reacting badly to the term “metadata”. In the linked entry, the rant is in the context of IT and enterprise resource planning (ERP), but I’ve also seen other vehement objections to the term (see the Wikipedia criticisms on it). This concerns me because it’s a word I’m using in my upcoming talk. I think it’s a useful word for capturing the idea of extended context about the original data (/information/object). If I want to save a reference to a paper I found online, and I think it’s a precursor to current discussion about applying game design principles to learning, I’d like to be able to tag it with “game” even if that word is not in the original paper.
“Metadata” seems like a good word to describe my use of “game” in that context. I don’t like the thought of turning off audience members with careless buzzword usage, but I’m not sure how else to capture what I’m getting at. “Extended context data” doesn’t flow so well for me. Also, a number of the criticisms of the term don’t seem to recognize how metadata can be cheap and fun (tagging/feedback), or how emergent metadata (e.g., from the general public using unstructured tagging) can lead to interesting informatics possibilities.
I’m going to stick with metadata … for now.
2 Responses to “Metadata, schmetadata … or useful context?”
I enjoyed your presentations at PenguiCon. That was my first con and it was a very good experience.
Anyway, most of the Wikipedia listed critiques of metadata seemed to miss the point. For example, the claims that metadata is expensive and that there is really no limit to how much metadata you could associated with a given dataset.
Now certainly, if we try to get an unlimited amount of metadata it will be very expensive, but most of us are only going to care about a small number of specific contexts which metadata can be used to describe. For example, most of us aren’t going to want a photo tagged with the names of everyone in a crowd at a soccer match, but if I know my daughter is in the picture, I’m going to tag it with her name. To say I’d then need/want to tag it with everyone else’s name is absurd (this seems a bit like saying that metadata is a bad idea because a picture of the Earth from space could potentially contain metadate for the names of everyone who has ever lived…these same like silly straw men claims).
What we call “metadata” is just one instance of a much broader categorizating that it’s hard to imagine we could do without. Long before we were adding metadata to essays describing game, for example, we have organizational systems for books that would divide them physically into a game category broadly and then in subcategories like baseball, basketball, etc.
Adding information in this way seems to be very fundamental to the way we form conceptual models about information.
The only real problem is the silly top-down enterprise solutions that turn it into a meaningless buzzword.
By Brian Carnell on Apr 27, 2007
Hi Brian, I’m very glad you enjoyed my presentations!
“What we call “metadata” is just one instance of a much broader categorizating that it’s hard to imagine we could do without.”
Sure. Categorizing is a survival behavior — is this one of those “dangerous” things I need to hide/run away from? Is this plant “poisonous”? Etc.
“Adding information in this way seems to be very fundamental to the way we form conceptual models about information.”
That’s an interesting way to think about modelling. Slightly different from the class-inheritance stuff I studied in Artificial Intelligence (old-school, ontologies for AIs, etc.).
By configures on Apr 27, 2007