(with apologies to Meg’s “What We’re Doing When We Blog”)… Discussions of “folksonomy” are meme-ing across the blog-space and I am disappointed that it is yet another round of issues being encamped in dichotomies.
I am picturing something like a “Meta Data Professional Wrestling Smackdown” (imagine a deep booming voice, not mine as no signs of any string of “ummmms”) bellowing:
IN THIS CORNER, weighing in at a taut 115 pounds, dressed in horn rimmed glasses and a bone chilling righteous stare, we have the CONTROLLED VOCABULARIAN, where the only meta data is well constructed, precise, and strictly defined by a coven!
IN THE OTHER CORNER, weighing in a a dripping wet 110 pounds, dressed in tie-dyed colors, cargo shorts, and 10 year old Birkenstocks, we have the FREE FORM SOCIAL TAGGER, where the wisdom of the crowd rises meta-data above all to a pure level of Waaaaaaaaaa.
Who will win? Or, does it matter? Why must it be either controlled vocabularies or folksonomics? Why not both?
In Liz Lawley’s “social consequences of social tagging“, she takes the vocabularian stand by pointing out the problems of tagging images via the ESP Game site— her criticisms as to why that site (where two players pitted against each other aim to match tags on the same image and thus win “points”) I totally agree with. But what I do not see is how it is is to her “the perfect example to illustrate [her] concerns”.
Why? Look at the motivation of the tagging. In the ESP game, the motivation is a “game”, to score points, and even more telling, the image being tagged has no ownership or importance to the person(s) doing the tagging. I think there is a huge gulf of that mode of taggin comprare to what people do when they tag their own photos in flickr or their own del.icio.us-ed web sites. The ESP game is a bit novel, but c’mom, does anyone really expect useful tags to arise from content the taggers do not care about?
When we tag our own stuff, we are doing it for a number of reasons- for our own organizational schemes, to perhaps add to a more global collection, to add to the wonderful serendipitous connections tagging opens up. Again, I do not think tagging your own content, or content that is important to you, compares to a web site game.
But my point is (I think) that we ought to be talking about the strengths of both approaches and where they work together. Take flickr. (No do not take it away!!) When you upload your images taken from digital camera, the image meta data recorded by the camera (image size, exposure settings, camera make, etc), is well structured meta data that is automatically added to the content. It works cleanly, and provides some hooks into perhaps analyzing images by this data. This works side by side with the loose tagging people apply on top of this.
See, controlled vocabularies working side by side with those sloppy tags.
Or take Google. I recall reading somewhere, that the technology for indexing and searching site is built around well controlled meta data (information about a URL such as its last update, a caches version, the title, description, etc). This well structured meta data is what enables the Google Magic to do what it does.
But when you look at how Google spiders the web, it pulls the information it searches from less then perfect structured content- sloppy HTML, and a slice of contence from the first X bytes of data in the page. And how about those META tags in the HTML <HEAD>…</HEAD> — there are web page designer applied keywords– more or less folksonomic tags.
See? Structured and tagged working together.
So can we by pass the trivial arguments about folksomonic tags replacing structured meta data? That is just ludicrous. They will always be chaotic, sloppy, localized, imprecise, and are never intended to replace what structured meta data can do. But they offer an amazingly easy and explosive way to connect disparate sources of content in fascinating ways. Do not lose that.
And while structured meta data provides the framework for setting up systems that are easily searched, shared, moved from one automated system to another, we ought to accept that asking mortal humans to repeatedly supply data for 11, 30, 80, 324 item data forms. I would rather do my taxes than fill out meta data forms. I would rather have a root canal. I would rather sit through a week of driving school courses (no, that might be going too far). Meta data ought to be done automatically, or near automatically, transparently, or by the professionals that like doing it.
My other gripe about tight structures is that they frame the perspectives and ideas of only the people who craft them– and leaving out completely the ideas of the people using he data or ideas that fall outside the rigid boxes.
So can’t we have both? Don’t we?