Great video talk from TED.com here by a guy called Blaise Aguera y Arcas  http://www.ted.com/index.php/talks/view/id/129 
He demonstrates his technology that takes something as complex as a whole book, and lets you zoom in closer and closer on any part of it from afar, and read it with the same clarity as if you hadn't had to zoom in on it at all. 

Then there's the example he gives of Notre Dame cathedral in Paris.
And the thousands of photos of the cathedral on Flickr.
They'd all been tagged 'Notre Dame', and Blaise's technology shows you a kind of montage of all these different images:
some taken in high res, some with mobile phones. And some were even photos of posters of the cathedral.
And again, you can zoom in on any individual part of the montage to reveal more and more detail about it.

It might be visual detail, or detail revealed in the tags that people have used to describe their images.
For example, the names of the saints who appear as statues on the cathedral's facade.
Blaise paints a picture where all the images of Notre Dame or whatever are linked to other images of the same thing, creating not only an incredibly rich metaverse (universe of tags) associated with that particular subject or image.
But also putting the observer in a position where they're one click away from all my other images, if I'm one of the contributors, thus creating a phenomenal network effect.

But you gotta see it!

The question that I asked myself was:
if we're able to take something as complex as a whole book, and zoom in on a very precise part of it in one movement, does this do away with the need to categorize information using knowledge trees? 
My hunch is that Blaise's technology will be great when you know exactly what you're looking for......when you know what part of the book you want to drill down to. 
What it won't be able to give you is a 'table of contents'-type overview of the content available, whether the content is text (like a book) or multimedia (like the cathedral). 

Especially as the volume and complexity of the content in question grows, my hunch is there will still be value in being able to have that kind of 'table of contents' view across a video or a collection of videos, for example, by browsing the tags (in the form of a knowledge tree like Windows Explorer with its folder and subfolder names) that point to the content

Have a look at the video and tell me what you think