In this section, we will explore another, different kind of preference tracking, called "collaborative filtering."

First, you must establish a username and a password at Amazon.com. This exercise works best if you have already bought a book from Amazon. If you have, they will have recommendations for you at the top of the page. If not, after you have established an account, Amazon will step you through the process of refining your recommendations

Spend some time—at least an hour—refining your recommendations. How accurate have the recommendations become?

"Collaborative filtering"works with massive databases such as Amazon's—which has tens of millions of people in it. When you buy or rate a book, the computer instantly checks for other people who liked or bought the same book. What other books did those people also like or buy? It then returns these items to you. The more you use it, the more accurate it becomes. It turns out that in a database of 20 million people, there will be many people whose tastes in books and music are almost identical to yours. How many? That's not altogether clear.

Some firms have attempted to use collaborative filtering as a kind of dating service, arguing that it should be possible to literally find one's "soul mate" through collaborative filtering.

Collaborative filtering is used when you purchase goods in stores or supermarkets. The data on your purchases is cross referenced with other people's purchases, so the store can compare you to people who like similar things and present you with you precise, targeted information on sales, or announcements of new products.

It seems reasonable to suppose that in the future, communities based on consumer preference will replace membership in nations or tribes—that you will belong to "the united colors of benetton" or to the community of Phish fans, or an international community of mountain bikers, or drivers of VW Jettas, or whatever your personal preferences may be. A quick Google search for "mountain bikers, association" turned up hundreds of sites, including the following:

MMBA : Michigan Mountain Biking Association www.mmba.org/
Mountain Bikers of Santa Cruz - Santa Cruz, Californiawww.mbosc.org/
The Maryland Association of Mountain Bike Operators www.bikemobile.com/mambo/
The Ohio Mountain Bike Association Website www.joinomba.org/
Middle Aged Mountain Bikers Association- Middle Aged Mountain www.ocnow.com/community/groups/Mamba/
San Diego Mountain Biking Association www.sdmba.com/
IMBA International www.imba.com/international/

This very brief list suggests the extent of the community of mountain bikers. These people share a set of interests—in certain consumer products, in certain activities; they share a set of political interests and beliefs about access to the natural world, about exercise, about health. They share far more than do the many people who are otherwise united under the flag of the United States.

Modern technology allows them to be targeted more precisely by advertisers, who in turn argue that people tend to form communities with people like themselves. Will such forms of marketing make it harder and harder to live with those who disagree with us? Harder to form "nations," which unite us not according to lifestyle preference, but according to political beliefs?

Questions to Consider:

• How does this process work?
• How accurate were the recommendations?
• Did they become more accurate as you used the system more?
• Could this system be applied to other things?

Answer these questions in your web journal.