20011114: Wurman, Finding things

Contact:cdent@burningchrome.com


Wurman, R.S. (2001). Chapter 8: Finding things. In _Information
     anxiety 2_ (p. 155-178). Indianapolis, IN: Que.

On Mon, 12 Nov 2001, Sonia Gupta wrote:

> I thought this article was interesting, but it alluded to the concept of
> Information Anxiety in only a few lines. And that too only for the case of
> search engines that either return too much or too little information (p.
> 172). I would have liked to know more about how we experience information
> anxiety in normal everyday life, other than while using search engines.

I think information anxiety is a misnomer, or perhaps a misdiagnosis.
Rao, going on about out roots, says, "It is in fact natural for us
humans to process large amounts of information." Dealing with search
engine results is an adaptation of an existing, innate, skill. If
someone finds that they can't deal with the results as given then
getting the results is not particulary important to them. The volume
of information that is accessible gives the impression that it must be
important by sheer mass, but it isn't really, is it?

Sometimes it is, and then we spend the time and energy to figure out
how to get to it. We "have the impetus to make it understandable".

Once we have the impetus, several things will happen: we'll learn to
deal with the results, and the methods for dealing with the results
(once we know what they are) will become more effective.

> Another thought regarding the 'Valence Project' being developed by Ben Fry
> to visualize text in 3 dimension (p. 161). I didn't understand the
> practical application of this project. How are they going to use the
> results of this project? Do they want to know which words are being used
> most frequently in books? And do they want to use that information to
> improve the quality of search engines? What other purposes can they use
> the results of this project for? Any thoughts?

I'm completely speculating, but word proximity and frequency analysis of
documents helps to indicate similarity between two documents. If you
map two valence images of documents you may be able to see
relationships or similarities. Clustering and overlaps that you didn't
know about.

The gzigzag project I mentioned on this list a while ago has an
interface that's quite similar to the valence interface.

I'm basically with everyone else: I think this particular chapter was
sort of a "hey, here's some cool stuff that is somehow related to
maps".

Information visualization is fully buzzword compliant out here at
Supercomputing. Every booth has some kind of demonstration of a neat
way to take huge volumes of info and stick it in a picture. IU's booth
has a display with a pixel density of 400 dots per inch. A usual
screen is 72 (if I'm remembering correctly). It's incredibly to look
at. It can be helpful in visualizing things like the human genome
which has so many pieces of information that normal screens simply
can't show it: there's not enough granularity.


Back to the Index