Monday, November 06, 2006

The trouble with tagging

Am I the only person who thinks that giving the user the ability to add 'tags' to the library catalogue is a bad idea? This thought isn't completely apropos of nothing. Folks are meeting to discuss the future of the Integrated Library System. I don't want 'tagging' in our future OPAC.

Here's why:

To me, tagging makes best sense in a context where words need to be added in order to be found. Photos (Flickr) and videos (YouTube) are just two instances where tagging makes it easy and rewarding for the user to annotate their own work because otherwise it would be unlikely that the piece would be found at all. But when you are tagging a piece of text that you've written, in essence you are either duplicating the words that you already used in your piece of text, or you are using more generalized words or synonyms of the main ideas in your piece of text. Essentially you are tagging a text with words that you didn't think were important to include in the text in the first place.

Now when we think of tagging in the library context, we are thinking of a user supplementing a text record (created by a professional) to make it more complete and findable. The idea is that, with tagging enabled the library catalogue becomes social entity, being improved upon by the wisdom of the crowds. But unlike projects like Wikipedia, there is no social reward for improving a catalogue through tagging. There's no incentive - no boost to the ego. And then there's the chance for vandalism.

I think my biggest concern about tagging is that there is the understanding that tags never expire. Once a record is tagged for someone's to read list, it will stay there forever. The idea is that every record will accumulate enough "correct" tags that the signal will overcome the noise. But that's not necessarily going to happen [via]. Over time, tags clouds inevitably amount to a collection of very general words and we know, as librarians that searching for generalized words like 'history' create very large and useless search results.

All this criticism aside, I don't want to take away from the work that the University of Pennsylvania Library has done in creating a library catalogue with tags. For one, they've demonstrated that the temptation to 'deface' a library catalogue with junk tags is largely unrealized. Besides, I think they've created a better tagging service than Amazon and there are very few instances when one can say that of our library systems.

2 comments:

Mita said...

Art: I agree that a large measure of the success of social software is how strong the 'social' part is behind it. I suspect that the best scenario for text-tagging is a small and active group who are dealing with shared and specialized interests and texts.

For example, I think unalog works because there's a small audience with similar, specialized interests. But remember when someone's young daughter tried unalog? All of a sudden there was a rash of posts about giraffes. (Or was lions? tigers? giraffes? giraffes!) In a larger social bookmark aggregator, these links would be just be background noise. But I would rather use unalog - with giraffes and all!

Mita said...

All of my misgivings about tagging in the catalogue have been made mute by the user-interface (in progress) of Bibliocommons. If you are at 2007 OLA Superconference, don't miss their demo!