Less than a year ago, Internet art was the flavor of the month, and no museum with an eye on its cutting-edge credibility could afford to ignore it. Witness the Data Dynamics show at the tech-friendly Whitney Museum of American Art and--at ground zero of the dotcom craze--the much-hyped 01.01.01 show at the San Francisco Museum of Modern Art. Unfortunately, Net art's criticism-proof cachet seems to have gone into receivership along with the rest of the New Economy. The novelty of Net art appears to be wearing off for curators at top-flight museums, and it's beginning to look a little peakish.

Unlike those collapsing IT companies, however, Internet art sticks around even when the public loses interest. Sites like Rhizome.org both archive older Net art exhibits and exhibit the work of people still honing their craft in digits and code. These sites make it possible to get a sense of the breadth of Net art and to find its most interesting avenues of development. With Net artists still busily trying to make sense of what, exactly, Net art is (as this e-mail exchange among a handful of Net art illuminati makes explicit), the most promising Net art works seem to be the ones which try to make art out of the structure of the Internet itself--that make us think not just about what the Internet is delivering, but how. Bluntly put, the most successful of these artworks are the useful ones--they don't just look good, but are practical as well.

iven its sudden ubiquity, it's easy and tempting to think of the Internet as a utility, like plumbing. Twist the spigot, out flows the information. Brian Judy's The Grid literalizes this simile, casting its audience in the role of the plumber. Judy describes The Grid as a "network of browser interface studies and interactive trinkets," but it looks a lot like a gradually expanding mess of pipes and valves. The computer responds to movements of the mouse--albeit somewhat unpredictably--by rearranging the pipes and changing the direction in which the system expands.

Its interesting conceit aside, as visual art The Grid is awkward and hopelessly incoherent. The images have the distinctive quality of a video-game background, and it serves up an empty, pointless version of interactivity. While it's not entirely clear that Judy intends The Grid to represent the Internet, the implication that it does so is quite strong. But the aimless wandering encouraged by The Grid is a poor metaphor for the way we use the Internet today (The New York Times recently reported that Web users do less surfing than they used to, and tend now to move directly to what they need). While it's tempting to imagine the Internet as nothing more than an impossibly vast network of connections, those connections are its means rather than its ends. The Internet as envisioned in The Grid is like plumbing that doesn't connect to the water main--not good for much.

Simon Biggs's Babel takes a different tack, imagining the Internet as a vast library. Something like a 21st-century card catalog, Babel uses the Dewey Decimal Classification (DDC) system to catalog the contents of a great many Internet sites. The mass of numbers thereby generated appears on-screen in the form of a nebulous cloud. As the cursor moves about the cloud, each classification number selected is displayed, along with the name of the website it indexes. Each of these relationships is individually determined by Biggs himself, according to the classifications codified in the DDC system. It's charming, in a way, especially since the cryptic numeric code of the DDC is convincingly computer-like.

Though Borges's imaginary infinite library is clearly an inspiration of Babel, his work resembles Borges's library only in that it won't help you find anything in particular. At root, Babel is little more than an elaborate demonstration that the Internet can be classified per the categories offered by the DDC. As comforting as it is to have that conclusively proven, there was never any reason to assume otherwise. The problem is that the DDC is not really a method by which one looks for a particular book, but rather a way of finding it on the shelf. Babel's mass of undifferentiated call numbers is impossible to navigate, and ultimately one is reduced to Borges's own Sisyphean method of randomly selecting one at a time. Its droll premise discounted, Babel is entirely artificial: This application of the DDC neither provides any apparent utility nor illuminates any special feature of the Internet.

he most successful instances of database-related Internet art are organized not according to some outside system like the Dewey Decimal System, but are rather more sensitive to existing patterns of the Net itself. To begin with, a natural symmetry is immediately established between form and content, which lends itself to a clean spareness rare in more-is-better world of Internet art. Just as important, they're a good deal easier and more pleasant to use, emphasizing an appealing utilitarian practicality.

For instance, while Lisa Jevbratt's 1:1 features a set of cryptic numbers used to index websites, in this case they are the Internet Protocol (IP) addresses that organize the entire World Wide Web. (Every website has an IP address consisting of a combination of four numbers ranging from one to 255 which allows your Web browser to find it--in the case of thenewrepublic.com, it's 209.143.193.110). 1:1 enables a user to navigate the Web via these IP addresses, organizing them in several different ways to illustrate both how they work and how they structure the Web. Particularly appealing is the interface titled "every," which provides a graphic representation of the IP addresses cataloged to date by translating the numbers into bands of color.

The title of the artwork, 1:1, refers to the scale of the map Jevbratt eventually wants to create. Unlike Web browsers, which Jevbratt likens to road maps, indicating only certain relevant features of a given landscape, 1:1 indexes every feature of the Web, providing a map identical in scale to its subject. To date, Jevbratt has indexed about 2 percent of the Web, or 186,100 IP addresses. What makes 1:1 so fascinating is the grace with which it illustrates a fundamental feature of its medium. As Jevbratt indicates, "The system is the art, not the output, not the visual screen, and not the code. I want to let the data express itself in the most beautiful possible way." Whereas Babel imposes an extrinsic system on the data it organizes, the system illustrated by 1:1 is intrinsic, something derived from its material rather than imposed on it.

Two other especially successful examples of Internet art, StarryNight by Alex Galloway and Mark Tribe and Spiral by Martin Wattenberg, use astronomical motifs to elegantly catalog all of the text files that have appeared on the Rhizome website. The simpler of the two, StarryNight, consists of a field of randomly distributed stars of varying size, the magnitude of each corresponding to the frequency with which the text it indexes has been viewed. Selecting a star brings up the option to read the indexed text as well as a list of keywords related to it, but no identifying title. Best of all, if a keyword is selected, a "constellation" consisting of all of the stars indexed by that keyword appears.

Whereas StarryNight organizes the texts with respect to their popularity, Spiral emphasizes their chronological relationship to one another. The stars are distributed in a spiral galaxy, those corresponding to the oldest texts placed closest to the center. By adjusting a slider on the right-hand side of the window, one can travel through time, causing the galaxy to expand or contract in order to display the desired period of time. Each arm of the galaxy represents a category (reviews, commentary, theory, etc.), so by traveling along an arm one can watch ideas develop. Both are remarkably efficient and logical means of archiving the texts, and quite lovely as well.

Painters, sculptors, even filmmakers, all have the luxury of a rich history that defines by example exactly what their medium does, providing the good, solid ground so conducive to innovation. If artists hope to make a medium out of the Internet, they're going to have to start laying that ground. Using art to represent, explain, and comment on the way the Internet works is the right place to start. If the Internet does turn out to be a promising medium for art--something that I think remains to be seen--it will be because of smart, humble, and diligent artworks like 1:1, Spiral, and StarryNight.

BRIAN FRYE is a filmmaker, curator, and writer living in New York City.