I remember the pain we went through in bringing HP's FTP site to the web. The first step was converting the old README files to HTML. Perl made this a snap, but somehow it didn't address the now-more-apparent quality problem. So we slurped it all into a Paradox database had a big metadata entry party.
Ok, the word "party" might be a stretch. There was technically pizza involved, but it was more of a bribe. It lasted days, and nobody really celebrated 'till it was over.
I distinctly remember the phrase "metadata monkey" entering my vernacular at that point.
I don't have anything against monkeys. Monkeys are cute. Monkeys at keyboards are even cuter. But metadata entry has long been viewed as a thankless job.
Now, sites like del.icio.us have figured out a way to get metadata monkeys to work for free. The incentive? Not pizza, not even bananas: just the ability to store our own descriptions, share those descriptions with others, and access it all from anywhere.
It makes me wonder about the role of the library in creating authoritative, versus curating social, metadata.
1 comment:
'Twas a librarian who pointed out to me that the nugget of Google's page rank strategy (determining what's significant about a URI by looking at the terms people use when linking to it) is a great example of getting reasonably accurate metadata from monkeys. Maybe the real issue is making sure you have enough monkeys ...
Post a Comment