David Weinberger has been blogging on the TTI Vanguard conference (OK I'm envious) and it looks like there are some voices being raised about the silliness of expecting people to behave like machines so machines can understand people.
Three clips.
People don't like filling out forms and entering metadata explicitly. (Or we refuse to do it, ignore it whenever possible, and do it perfunctorily the rest of the time) So, a KM system ought to mine content for metadata. (Or mine relationships, remember, Google doesn't know anything about anything, only that people think this URL is more important than that one. Content may be just a token, a beacon that identifies the location of knowledge, like a lighthouse on a rock)
Filling in metadata makes us pull back from the world, an attitude that goes against our biology. In fact, it's desire itself that draws us into the world and makes us shudder as we draw back from it. (Mhmm, looks like a very interesting conference)
The way people describe how they share information isn't how they share information. (The way people describe how they do anything, isn't how they do it, that's why usability testing doesn't include surveys, we either lie, or we don't know so we guess.)
I keep banging on about Google being a reputation tool, I want to revise that. I think PageRank measures consensus even when there is no agreement. Agreement is a negotiated property, even when we agree to disagree, we have to negotiate it, but consensus is an emergent property. Even in face to face meetings, a group of us will talk, maybe for a long time, and then someone will say "I think we have a consensus here" and then we will agree or disagree that it is so.
That's what Google does, listens to a whole lot of people talking and then says "these people have reached a consensus about this subject" and if enough of them are included in the consensus, the URL rises to the top.
OK, probably trivial.
Comments