LIS 9763

Quintarelli states folksonomies “are not the solution to every modern problem of classification”, nor are they an alternative to traditional classification schemes; rather they have their own merit as a powerful tool with special properties. After reading the articles and playing around with a few of the sample systems, I have to agree: for sure, folksonomies can be useful and offer have a lot of possibilities, but I don’t see them replacing traditional cataloging practices anytime soon.

There’s no question that folksonomies have a lot to offer the world information management. For starters, folksonomies are fast: because there are no controls or protocol, the language can change as soon as it evolves. As Kroski rightly points out, folksonomies are current, and in the digital age this is a feature cannot be ignored. It seems every day companies are offering a new product, developers design a new device or type of code, and new slang becomes commonplace. By using a folksonomy it is possible to keep up with these changes immediately, while eliminating the guesswork: authors tag and organize information in ways that are meaningful to them, as soon as the content is published. For the searcher, folksonomies offer multiple paths to the same piece of information. Say, for example, a user decides they wish to find a recipe for dessert: they can find different recipes depending on their mindset and what they enter. A search for cheesecake, pie, cooking, baking, dessert, recipes and culinary arts may all brings back viable results. As I hinted last week in social tagging, when it comes to browsing, folksonomies are excellent, giving the searcher a good glimpse of what’s out there. Both the Blais and Danbury cataloges give a good example of this: click on any of the LibraryThing tags and the user will see a list of books the library has that are also given this tag, a potentially useful feature were I a patron of either facility.

The catch-22 about folksonomies, is that while they are efficient at bringing a selection of possible results back to the user, they cannot bring back everything on an assigned topic, nor can the results be narrowed down to find exact documents. Suddenly, the folksonomy’s strength is its weakness: because the user can tag with different words that are meaningful to them, there is not any one word that is used on all items of a given topic. As many authors, including Mathes note, there’s no synonym control. A search for “bunnies” for example, will ignore results if tagged with “bunny”, “rabbit” or “hare”. In addition, when we tag items in ways that are meaningful to us, we need to ask: what happens when someone with a different mindset comes looking for information? It may make perfect sense for one user to tag a list of market meats as “dinner_ideas”, but a vegetarian probably wouldn’t agree.

Allowing the user to define where items are classified also present a different sort of problem that folksonomies inevitably need to deal with: what’s to stop a seller from tagging their advertisement or product with every word in the dictionary, a la spamming, just to ensure it’ll come up in the results? Worse, what about X-rated content being tagged with seemingly innocent words that young users might enter, such as “cute” or “sweet”? Lawley hints at this, although Technorati seems to have updated their censorship between her and a college’s search. Flickr forces users to tag with “safe” “moderate” or “restricted”, however there’s no guarantee every tagging service will have this as part of their policy, and even if they do there’s no guarantee a user won’t tag an item “safe” when it should be “restricted”. I’m not an advocate of censorship on the web, nor do I feel it is possible to avoid all offensive material given everything that’s out there, but if the library advocates a social tagging system, particularly for content beyond their control, I would pause to ensure the system has safeguards and a limitation of liability notice.