More on TechnologyTell: Gadget News | Apple News
‘Cory Doctorow: taxonomies/metadata: it’s all crap’
May 3, 2007 | 1:01 am
|
By David Rothman
“Well, the problem with explicit metadata, as I sum up in the essay, is manifold. But it’s that people lie; they tell you what they think you want to hear. Or, they tell you what they think they believe, even if it’s not what they actually believe. People are dumb, right? They sometimes just have bad classification information. People are lazy, so they misclassify because they can’t be bothered to properly classify.” – Cory Doctorow, as quoted in Wired News.
The TeleRead take: How true! Tags can be so bleepin’ subjective. I’m including this item under “e-books” in various variants, but some might want to restrict it only to “libraries” and all that.
Related: Everything Is Miscellaneous, David Weinberger’s new book.
|
Previous
SUBSCRIBE TO RSS
Comments:
The late Kurt Vonnegut had a chapter in his book _Cat’s Cradle_ titled “Never Index your own book”, which was funny enough that it often comes to mind when I think about the perils of self-classification.
The interview you link to is interesting for a number of points. one of them being the importance of considering use cases when you’re considering what role metadata should play in your system. The “good enough example” principle is acceptable in many cases, such as in searching for a decent photo of Capri. But there are other use cases, including a number in academic research, where you want more than “good enough”; you want to be sure you have a fairly comprehensive picture of reviewed materials in certain areas. Here, explicit metadata schemes, with the right tools, can play a very important role.
One does need to have appropriate tools (and many tools that use explicit metadata, such as most online library catalogs, are too poorly designed to use them very effectively), and one needs to recognize that implicit metadata can be useful too. In some ways, the data mining tools we’ve developed for implicit or uncontrolled metadata are more sophisticated than those that are in common use for explicit, controlled, metadata, which may lead to the mistaken impression that explicit metadata is useless.