Netscape's "cookies" can be used to destroy privacy on the Web or to enhance it. Cookies can be a force for good or for evil. Unfortunately, the choice is not ours. Or is it?

Netscape introduced its cookie specification with Navigator 2.0. Simply put, a cookie is a little tar ball of data and bits that a Web server can throw into the copy of Netscape Navigator (and many other Web browsers) on your hard drive. Once you've got it, every time you click on the Web site, your browser throws the cookie right back.

Cookies have taken a lot of heat since they were introduced earlier this year. That's because they seem to remove one of the great features (or problems) of the Web: anonymity. Cookies make it possible to track a user's movements on the Web. And since cookies can be persistent, staying on the user's hard disk for months or even years, they make it easy to see how a person's interests or use of a site changes over time - an invaluable tool for marketers, in particular.

But what's really unsettling is that everybody's getting into cookies - even sites with nothing to sell - just to track Internet users. Most privacy activists whom I know abhor such cookies. They've set their Netscape Navigator preferences to warn them when cookies are received.

Despite frequent diatribes, I'm a big fan of cookies. Used properly, they can actually increase privacy and improve the Web experience. Cookies only violate a person's privacy when they're used to track a user's movements and index them in a central databank.

But you can also improve privacy by using cookies to eliminate the need for a central databank. That's especially important for Web sites that are looking for ways to offer customizable interfaces and individually tailored content delivery. Using cookies, these services can be offered without storing lots of personal information for each subscriber on the Web site's master servers, which is exactly what makes privacy advocates nervous.

The secret is to store a user's preferences in the cookie itself. For example, a Web site might download a cookie to a person's Web browser that records whether the person prefers to see Web pages with a red background or a blue background. A Web site that offers news, sports, and financial information could use a cookie to store the user's preferred front page.

A few months ago, Lincoln Stein published an article in The Perl Journal that demonstrated how to use cookies in this manner. Take a look at his demonstration program.

Unfortunately, using cookies this way takes a lot of work and thoughtful programming. For one thing, every time you change the format of your cookie, you'll need to be sure you can still read the old versions. It's a lot easier just to hurl a cookie at somebody's browser with a unique ID and then index that number to a relational database sitting on your server.

Keeping state in the cookie, rather than on the Web server's database, means you don't have to track sessions: Your server can become essentially stateless. And you don't have to worry about expiring out the database entries for people who clicked on your Web site six months ago and haven't been heard from since.

However, there's a problem. Web sites that store a lot of personalized information in your browser's cookie file - in the interest of protecting your privacy - will end up having to use fancy data compression techniques in order to keep the cookies from getting too big. And it will be impossible to tell those cookies from cookies that are merely tracking you in a database.

The Electronic Frontier Foundation thinks it has a solution to this problem. Called eTRUST, the program's goal is to develop standards for online privacy. One of the things those standards would govern is what Web sites can do with personal information they collect about their users. Web sites would display a particular eTRUST logo indicating their privacy policy; in return, they would submit to data audits by a recognized accounting firm.