The Beaver.

Seeing as how i've lived in North America for about 2 years now, I feel I am an armchair expert on the beaver. Plus, I have the internet to give me thousands of unqualified references to look at and refer to, building a case for my new-found expertise. Given that I am inclined to borrow knowledge from completely unrelated areas of study and squish them into whatever I am thinking about at the time, here is a another long shot post about tags, information ecology, ecosystems, and Beaver as Expert.

When beavers dam a stream they slow the movement of water. Behind the beaver dam, a pond of still water is formed. This pond (impoundment) is then colonized by animals and plants that typically live in lakes rather than streams. Organisms dependent on fast moving water die out in the beaver pond, or move to parts of the stream where the flow of water has not been slowed by the beaver dam. After a beaver dam has existed for ten years or more, the pond it created usually has an abundance of submersed and emergent vegetation, along with the many animals that live in such vegetation.

Ecology of the Beaver: Its Dam-Building and Tree-Felling Impact on Other Wildlife

Flickr displays many, many photos now. One of the ways people can look through them is surfing by tags. People create tags; sometimes on the spur of the moment, and sometimes through pre-meditated action. People copy tags, and assign them to their own photos and the photos of others.

To me, this sort of maintenance is equitable with the processes that occur within a natural ecosystem. This is where the beaver comes in.

Consider this:

When beavers dam a stream they slow the movement of water. Behind the beaver dam, a pond of still water is formed. This pond (impoundment) is then colonized by animals and plants that typically live in lakes rather than streams [who all play nicely together].

And then this:

When Flickr members tag their photos they slow the movement of the photos. Behind the tag, a pond of photos is formed. This pond is then colonised by other photos tagged with the same thing [and new, unique, emergent connections to other photos are made by tags related to, or connected with the tag that creates this particular pond].

It's the "submersed and emergent vegetation" that i'm interested in. There's talk of placing a semantic structure on top of the way people tag their photos. I think this is a bad idea, for many reasons. Any "structure from above" will be wrong for some people, not everyone will use it, and in my opinion it will kill the emergent beauty of what's happening on Flickr. The transformative power of me seeing a tag I like and using it on my own photos is working now, and working well. I can create my own individual tags that suit my individual collection, or I can choose to add my photos to the collective, emergent whole, by using tags that I see around the system.

I'm just not sure that any sort of impressed, expert classification system would be useful, and i'm certainly yet to hear from anyone about what might be sufficiently generic and lightweight, yet useful enough to be a credible proposition.

(Thanks to Kent Runge for the photo.)

Wednesday, March 8, 2006: In hindsight, I think I misunderstood "semantic structure" to mean an enforcement of something like the Dewey-Decimal System, which certainly has a use of sorts, but is a hopelessly generic system for findability. Perhaps all that's needed is a shared syntax for adding structured metadata. And structured data is good (in addition to fertile, messy, freeform, performative tagging, because it makes things even easier to find, parse and exchange. If people can create their own 'mess' within a syntax, i'd be happy with that.