And the tagging obsession continues. Thanks to Travis Smith for pointing me towards Larry Borsato’s comments on why we don’t tag our desktop. His post is a response to Kevin Briody’s call to tag your desktop. Kevin asks:
Why can’t we tag documents? And file shares? And intranet sites? Then tag communications: emails, Messenger contacts, and address book contacts?
Borsato thinks he has the answer:
The reason desktop search is so useful and necessary is because it helps us find the stuff we lost. Any usable system will required both the ability to categorize, and the ability to search across categories. The folders we are already comfortable with address all of Kevin’s requirements. They can aggregate documents, contacts, other folders, and links or shortcuts to other information.
But Larry is missing the three laws of tagging that make it not just desirable, but indispensable to the future of the desktop:
- Choices suck.
I’m often working on documents or files that belong in more than one place in the haphazard taxonomy that is my computer. I save a file to whichever folder seems to make sense at the time, but that’s no guarantee I’ll be able to guess that location when I go looking. (What can I say? I am mysterious and enigmatic, even to myself). And since today’s long filenames still impose some choices about what words to include or leave out of a filename, you can’t count on search to bail you out. With tags, you don’t have to make those tough choices…just slap on all the keywords you think you might ever need. - Flat is better than pointy.
The beauty of tagging is that it’s a totally flat system…no nested folders or hierarchical categories to worry about. Folder-based desktops, on the other hand, are very very pointy, digging deep into the bowels of your hard drive and spiking up here and there as top-level folders. And like any pointy thing, they require a lot of pruning. I for one am sick of continually pruning my desktop folders back into some kind of coherent order. - Many is better than one.
And if it isn’t enough that tagging can save you from your own leaky memory and desktop disarray, think about what it can do when combined with the rest of the big world. Anyone who has checked out a Technorati tag page has discovered the joys of converged tagging. Just imagine being able to combine all the junk roaming around your desktop with all the related material on del.icio.us, furl, flickr et al.
In fact, why stop at the desktop? I want a labelmaker that will generate text/UPC tags so I can tag objects around them house and retrieve them using our in-home laser-guided Universal Locator System. Oh wait…that’s the future. Why isn’t it here yet?
Hey a little hierarchy never hurt anyone. And how could I possibly know all the tags I’ll ever need. 🙂
The truth is I get a little tired of tags as “the next big thing” that will change technology as we know it that will solve all of the problems we have now and forever. You would no sooner have one big desktop full of thousands of files with tags, than you would have thousands of individual sheets of paper with post-it notes on the on the floor of your office.
To be efficient we need some combination of hierarchy and tagging, and search is still very useful.
And where purely text data is involved, tags reduce the value of the information. A picture may be worth a thousand words, but a tag isn’t.
I’m a big proponent of using something like latent semantic analysis to generate tags automatically from my text.
That said, I think tags are great for things like del.icio.us and Flickr.
Mostly I just love to get a goo discussion going.
“I want a labelmaker that will generate text/UPC tags so I can tag objects around them house and retrieve them using our in-home laser-guided Universal Locator System. Oh wait…that’s the future. Why isn’t it here yet?”
Actually, it it here now. UPC tags that resolve to URLs via cameraphone are the brainwave behind semacode.com