Tagging is hard work when you're lazy
Posted on May 11th, 2005
I don't know about this tagging business. Back in September I wrote about not getting the appeal of del.icio.us because of the underlying effort required by any sort of manual annotation. Since writing that post I've been using Spurl, to manage my bookmarks. I've accumulated hundreds of them. I've tried to be diligent about tagging each bookmark and building a sensible folder hierarchy, but I'd say that 9 times out of 10 I bookmark something because I'm too lazy or preoccupied to read or process it at the moment. So my bookmark collection is both a collection of useful resources as well as a collection of "yeah, I want to go back to that eventually" resources. I dump the latter in a folder called "incoming." There's far more stuff in that folder than in any of the others. Realistically, I'll probably never muster the discipline to sift through everything in there and discard or retag and recategorize as appropriate. The more resources I collect, in fact, the more I consider tossing out the whole mess and putting my faith in Google to bring me back to the right one at the right time.
I read Peter Merholz's Metadata for the Masses article the other day, and I can see all the benefits that tagging provides. Tagging is really useful after it's been done. If Google or Yahoo implemented a tagging system, I think it would definitely add something positive to the search process. That's the appeal of having local collections in the first place, whether it's bookmarks or photos or music-- faster identification of the meaningful and relevant from everything else. When the collection gets big or moves off your hard drive and onto a shared service like Flickr, tagging helps with identification but it also helps with ownership-- I liked this resource so much that I decided tag it, and now the knowledge that it's there and that I can refer back to it when I want it is mine, all mine!
But tagging is still work that I'd rather not do. Tagging is extra effort on top of whatever else it is you're doing-- bookmarking a URL or uploading a photo, for instance. Tagging is entirely separate from that primary task. It's like flossing-- something that you do because in the back of your mind you know that's it's good for you in the long term. Better to tag now and increase the resource's findability later on. That's the part that I'm not sold on. I am admittedly still skewed toward the predictions of the future in The Age of Spiritual Machines, but even so, shouldn't the computer figure out what the right tags are? And then correlate them as needed with other tags, such that "nyc", "new york", and "new york city" are aliases for the same thing? Easier said than done with pictures and non-textual resources, sure. But would it be that big a leap to download a set of photos from your camera, then type (or speak, technology willing) a description of what they represent in a freeform, blog-entry sort of way, and then let the computer go to town with that plus whatever metadata it can extract by itself from the files (date, for example, maybe even color analysis... i.e. show me the greenest photos that are mostly likely forest pictures.