It's the agents, stupid!

With technorati adding tags in the style of del.icio.us and flickr, the whole 'sphere is abuzz with tag-talk. People talk about social consequences, regulations, stability and stamina of tagging by the great unwashed. I'm not linking it all up because there's a lot out there, but thankfully you already have! (del.icio.us tag: metadata, del.icio.us tag: folksonomy, del.icio.us tag: folksonomies)

I've heard horrible horrible neologisms such as 'folksonomies', and stabs at formalisation like 'flat namespace', 'metadata ecologies', 'grassroots classification', 'faceted classification', the list goes on. I've resorted to calling it MMOCS or Massively Multiplayer Online Classification System, which doc paradox assures me is pronounced 'mummocks'. At least it sounds funny.

Back to the topic at hand, I'm convinced that this is a problem to solve with an agent. Treat flickr, technorati and del.icio.us as sources of information and let a smarter piece of software guide you through the sea of raw data. Visualize a limited dataset real-time (based on one or more initial keywords), and show options for adding related tags, other-language-same-word tags. Make it able to toggle between your tags, other tags, popular tags, recent tags, tags from contacts. Turn input from selected services (flickr et al) on and off with a button. Mix it all up, and see where you end up. Maybe use flash or java3d to visualize it all realtime in a 3d environment, a great fractal webtree for people to explore.

I was reading the blog of Tim Bray, technology director at Sun, and from his blog post I would like to quote:

I’m pretty sure this is a bug: if I do a tag search for “Microsoft”, and if Robert Scoble or Jonathan Schwartz have posted on the subject any time in the last 48 hours, I want those at the top of the damn list, not what some college kid wrote three hours ago.

Now unlike Tim I couldn't give a rat's ass what Scoble or Schwartz have to say about Microsoft. So it's probably not a bug, just a personal preference on Tim's part. It's a problem of filtering, he just needs an interface that provides multiple filters tag: microsoft, author (or domain): scoble OR schwartz, time: within last 48 hours. Software can solve this. del.icio.us doesn't do it (it's not meant to), but I'm pretty sure someone is messing with something like this right now. People are writing little addons and derivatives of del.icio.us like there's no tomorrow, see here. (hmmm, self-referencing del.icio.us link!)

Another quote:

Should I be using my own category tree, or should I switch to tags that Technorati already knows about? Should I introduce Tim Bray and ongoing tags, presumably on all entries? Should categories be in trees? Should their names be monocased (with potential for i18n breakage)?

Without wanting to sound like a total hippie: Relax. Just tag away. If you spend an hour thinking on what you should tag your post as then you've just wasted an hour. Try to think up three tags that might have something to do with the subject at hand. Use more or less tags as needed. In my discussion with doc P it came up that she uses underscores to link multiple words together (as in this_is_a_tag), while I remove any spaces (thisisatag). Other people might use different symbols. Also, since I am dutch I might use my own language. To tag a picture of a flower on flickr I could use 'bloem' as a tag. A much larger crowd will use 'flower', of course. If the spanish-speaking crowd adopts these tools en masse you might see a huge shift toward spanish. Your agent software should take all these variations as well as plural forms of the words used for tags into account.

Easy? no. Impossible? no. Personally, I'm convinced this informal classification by groups of people is a great thing, and we'll see powerful tools popping up that can manage the vast 'flat namespaces' people are creating. Something in Java or Flash, and possibly orange in color.

At the very least I'm having a lot of fun with it ;)

Tagged metadata, tag, tagging, flickr, del.icio.us, technorati ... and why not?