Thursday, April 3, 2008

Folksonomies

Welcome back to Thesaurus Thursdays! Some of you may be wondering about how this Disneyland thesaurus relates to folksonomies, which are the combined result of people (like you, over there) tagging web documents. The benefits are the cost (hey, people do it on their own time--it's virtually free!), the closeness of the tags to how people see and think about things (they use their own vocabulary rather than one imposed upon them by a thesaurus), and the speed at which new concepts can be introduced to index documents. I certainly think folksonomies can be very powerful tools and a nifty aggregation of collective wisdom. But the purposes behind this Disneyland project don't mesh so well with the limitations of folksonomies.

First of all, a chief purpose for a thesaurus is terminological control (related to nomenclature, but not exactly the same thing). You want to use "Indy" to tag documents related to the Indiana Jones Adventure, I want to use "Indiana Jones Adventure: Temple of the Forbidden Eye." If you wanted to find all photos on Flickr, you'd have to try a number of search strategies and know every possible synonym somebody might use. That's using different terms for the same thing. You can also use the same term ("Indiana Jones") and mean different things (the movie, the character, the Audio-Animatronics figure, the attraction...). With a folksonomy, it's a bit of a free-for-all. And with Disney, there's never been a shortage of official names for things, so I'm not creating a bunch of my own terms. The language that Disney uses serves as the guide and should be pretty familiar.

A really big drawback for me regarding folksonomies is the lack of hierarchy. Davelandblog uses tags pretty well, but there's only so much one can do with a flat structure. For instaince, in a recent post on Matterhorn Climbers, the blog has tags for "Disneyland," "Fantasyland," "Matterhorn," and "Matterhorn Climbers." A good hierarchy would allow you to just have "Matterhorn Climbers"--all the rest is known from the structure. Then if somebody said, "Let me see all of Dave's posts about Disneyland," they could do an inclusive search and find that. Or if they just wanted to search on Matterhorn Mountain (inclusive), it would turn up photos of the mountain, the attraction, and climbers. I do not know of any structured folksonomies.

I believe most folksonomies that exist are pretty general. That is, on Flickr or del.icio.us, the whole breadth of human activity and knowledge is game for tagging. It's very difficult to develop an organizational structure for that sort of thing (and then to maintain it--!). But I do consider the history of the Disneyland Resort to be knowable and structurable. While it's certainly a significant task to assemble and maintain something approaching a complete record of Disneyland's history, it's only a few hundred acres and its story only occupies a bit more than fifty years. That's not a lot compared to everything (including Disneyland!). And it certainly seems more doable than trying to learn about every place in California (I only got up to around 12,000 before the thesaurus project distracted me).

Finally, the entirety of folksonomies seems to be the tags and the documents to which they are applied. There is no further metadata for the tags. My Disneyland thesaurus project is not just an authority file of index terms and lead-in terms, but contains a significant amount of information for the preferred terms. (Indeed, it is the basis for the encyclopedia.) In addition to relationships between terms, there are many different types of note fields, to record operational or birth and death dates, scope of sales or food products served, and especially definitions and other information gathered from various sources. I had tried organizing this in a Microsoft Word document in the early 2000s that grew to be over 1,000 pages, but found it difficult to think about all the relationships this way. Kevin and I believe that our ability to go through the thesaurus term-by-term will keep us from overlooking entries for the encyclopedia.

(It occurs to me after I write the above that I have included some non-folksonomy examples in my discussion here. Folksonomies are more properly when different people are tagging the same things, rather than when people are tagging their own things in a relatively unstructured manner. The limitations are fairly similar for both, however.)

2 comments:

Major Pepperidge said...

Folksonomies, there's a new word for me!

Wow, you are a serious compiler and organizer!

I wish I could be as organized as Daveland when it comes to implementing the tags. But I admit it, I'm lazy...!

Richard Harrison said...

Hi, Progressland; glad to have you as part of the Disney blogosphere!

As far as taxonomies gomy tags are all based on location, location, location. As part of the process of naming all of my Disney park music, I took the approach of existing scientific taxonomies: start big and work my way down.

This approach carried over to my blog posts and works something like this:
Park Name > Land Name > Attraction Name > Other Detail

As such, a post on the Matterhorn would be:
Disneyland > Magic Kingdom > Fantasyland > Matterhorn. (The Magic Kingdom tag is to separate from DCA.)

A post on Epcot's Living with the Land would be:
WDW > Epcot > Future World > The Land Pavilion > Living with the Land.

This helps me keep things straight but your approach of folksonomies addresses a much more important aspect of tagging - how to keep things relevant and searchable!