Every time a new technology creates legal problems, we face in a particular context the general question of relative institutional competence. Do we turn first to the judiciary, allowing time for a gradual solution derived from common law methods, or do we look instead to the federal legislature for an instant global solution? This Article endorses the judicial approach, suggesting that Congress is particularly likely to err when rapidly changing technology creates a perceived crisis, and when the strongest reasons for not legislating are abstract and inchoate. The Article examines three legal questions raised by computer technology, two the subject of recently enacted federal statutes and the third dealt with solely by the judiciary. The author concludes that both of the factors that maximize the potential for legislative error are present in most cyberlaw questions, and that the judiciary has therefore been more successful than Congress at responding to the legal problems arising from this new technology.

INTRODUCTION

Speed is an asset in computer technology, but not necessarily in law. The new technologies of the twentieth and twenty-first centuries have inevitably raised new legal questions; all too often, the response to these new legal challenges is a hastily enacted federal statute. If the Internet allows children access to pornography, we enact the Communications Decency Act ("CDA").1 Commercial concerns about cyber-authenticity prompt the Electronic Signatures in Global and National Commerce Act ("E-SIGN").2 Are cybersquatters creating domain name problems? We've got a law for that, too.3 These are just a few of the quick fixes driven by a perceived need to address new legal problems arising from new technologies.4

But many of these perceived "crises" could actually have been resolved by previously established law. The common law and judicial interpretation of existing statutes were working towards a solution-or had reached one-when the new statute threw a monkey wrench into the works. Each of these federal statutes has serious flaws, some of which arise because the statute short-circuited the common law's traditional method of dealing with new problems. The Communications Decency Act has already fallen;5 in its haste, Congress failed to appreciate the statute's constitutional implications. This Article argues that both E-SIGN and the Anticybersquatting Consumer Protection Act (ACPA), while not unconstitutional, are similarly shortsighted. More broadly, I suggest that, at least in the context of new computer technology, allowing time for incremental judicial responses is often superior to instant legislative solutions of a global nature.

When faced with new technology raising novel legal questions, we have two choices: we can wait for the courts to apply existing law-and eventually use legislation to fill any gaps that remain or to make new policy choices-or we can legislate immediately. Sometimes an immediate legislative response is appropriate, but

often it is not. We should be especially wary of hasty legislative responses in two circumstances.

First, legislation designed to address questions raised by a rapidly changing technology is likely to become obsolete equally quickly.6 Of course, judicial opinions on the same technology also become out-of-date. But once Congress has enacted legislation regulating a particular subject, it is unlikely to revisit the issue for some time; previous enactments can have unintended consequences or become straitjackets in changed circumstances. Judges are constantly reevaluating doctrine, however, tweaking it to take into account new factual contexts. Moreover, judicial opinions do not have the global reach of a congressional enactment: if one judge, or even a dozen, get it wrong, the geographic reach of the mistake is limited, and in any event new cases and new judges stand ready to make corrections.7

As an example, consider the American experience with the railroads. Like cyber-technology, the railroads stretched the law and tested its ability to address both new and developing issues. New questions arose across the legal spectrum, touching everything from employment law to tort law and from federal-state relations to anticompetitive behavior. The courts struggled for decades, restructuring and reconsidering the law to accommodate these challenges.8 Ultimately, the changes in the law were both successful and longlasting.

One of the most successful judicial innovations involved railroad bankruptcies. Beginning in the 1870s, railroads affected by economic downturns began to become insolvent. Since there was not yet any national bankruptcy legislation that would permit reorganization, insolvency might have led to dissolution and sale, disrupting national transportation on a massive scale. The courts, without congressional assistance, turned to--some would say "stretched"-the doctrine of equity receiverships to avoid that re

sult.9 Congress eventually modeled general federal bankruptcy reorganization legislation on these early judicial experiments.10

Direct congressional intervention was less successful. Between the Hepburn Act in 1906 and the Transportation Act in 1920, Congress passed a series of laws regulating railroad rates. These Progressive Era laws were designed to keep railroads from using their monopoly power over domestic transportation to unreasonably raise rates.11 Had the railroads maintained that monopoly, the statutes might have worked well. Unfortunately for Congress, however, the 1920s saw the rise of the automobile and the trucking industry, and the consequent demise of the railroads' monopoly.12 Hemmed in by a regulatory scheme that did not affect its roadway competitors, the railroad industry began to suffer, culminating in massive railroad bankruptcies in the 1970s. It was only those bankruptcies, some fifty years after the Transportation Act, that led Congress to revisit the question of railroad rate regulation and to give railroads more power to set their own rates.13

Thus, in the railroad context, Congress was at its best when-as in the insolvency context-it waited for common law processes to bear fruit, using decades of judicial experience to inform legislative enactments. The well-motivated but ultimately harmful legislation in the ratemaking context, on the other hand, shows the pitfalls of congressional intervention in a fast-changing technological environment.

A second factor that tends to produce flawed legislation is the lack of a concrete, defensible interest against the legislation. Here we might consider the recently invalidated Violence Against Women Act ("VAWA").14 Who could be for violence against women? This absence of competing interest groups is deeper than simply that the proposed legislation might affect a diffuse or unorganized or unmotivated group, and that the group therefore has insufficient

lobbying power.15 (No one worries about the unrepresented "interests" of people who commit violence against women.) The problem with statutes like VAWA is that the true considerations against the legislation are not susceptible to interest group politics, but rather fall to a deliberative Congress to consider.16 In the case of VAWA, for example, the interests on the other side were purely abstract and theoretical: the avoidance of duplicative legislation and whatever federalism concerns one might find legitimate.17 Moreover, because public opposition to such a statute is negligible, quick action is likely to be repaid with public approval, further increasing the likelihood of drafting mistakes. The lack of opposition also makes it easier for Congress to spend its time on these "easy" statutes rather than grappling with more difficult or controversial questions; if the federal legislature continually reacts to the crisis of the moment, it may neglect more long-term problems.18

Much cyberspace legislation, including the two statutes considered in this Article, triggers both of these concerns. Technological developments make regulation obsolete almost from the moment of its enactment, highlighting the advantage of the judicial ability to constantly reconsider and revise doctrine. Moreover, the speed of the technology exacerbates the crisis mentality of the public and other interest groups, putting great pressure on Congress to act now, while the reasons to postpone legislation are highly abstract. One such interest is avoiding the inevitable disruption caused by new statutes enacted against a background of developing judicial doctrine: the courts have been struggling with cyberlaw questions

for several years now, and each new statute creates new issues and new confusion. With legislation governing cyberspace content-from rules on domain names to rules on copying-there is also a potential First Amendment concern. Finally, there is the possible, if ephemeral, harm to the technology itself: an ill-advised statute might prevent the technology from developing in what might prove to be fruitful ways. In the case of the ACPA, moreover, there is an additional factor that increases the pressure on Congress. International harmonization is often used as a justification for laws regulating the global Internet, and multinational companies are eager to have Congress achieve harmonization. Since many of the countries that are legislating in these technological areas are civil law, rather than common law, jurisdictions, there is a further bias against judicial problem-solving.

Congress may well conclude, independent of public sentiment, that the benefits of the statute outweigh the harms to intangible interests. There may also be circumstances in which the costs of waiting through the long common law development process are greater than the risks arising from possible congressional errors. One example of such a circumstance is if the adoption of a standard-any standard-is more important than its particular content (such as determining which side of the road to drive on). However, we should be somewhat suspicious of quick congressional action where the two concerns identified in this Article are present. And, in fact, this Article suggests that recent cyberspace legislation justifies the suspicion.

This is not to suggest that cyberspace should remain unregulated. Scholars debating cyberspace regulation have, by and large, fallen into three groups. The first group, the cyber-anarchists, deny that traditional forms of law should play any role at all, suggesting instead that community norms and technological limits should be the only regulators.19 A second group, the cyber-exceptionalists, re

ject anarchy but nevertheless deny that traditional legal tools are adequate to regulate cyberspace; they urge instead the creation of a separate legal regime for cyberspace. Their model seems to suggest cyberspace either as a kind of separate sovereign (as the states and the federal government are separate sovereigns) or the subject of an independent body of law.20 A third group complains that both the cyber-anarchists and the cyber-exceptionalists exaggerate the novelty of the legal questions raised by the new technologies. Frank Easterbrook argues, for example, that developing specialized doctrines to deal with the legal problems of cyberspace is analogous to creating a "Law of the Horse." He suggests instead that we should "[d]evelop a sound law of intellectual property, then apply it to computer networks."21

I believe that the truth lies somewhere in between. The cyber-anarchists and cyber-exceptionalists have overstated the novelty of cyberspace, but Easterbrook and his fellows have undestated it. We cannot simply plug the factual context of the Internet into existing law and expect it to produce coherent answers, but neither do we need to create an entirely new regime to govern cyberspace. So how should we regulate this hybrid creature of familiar legal questions in a strikingly new context? Lawrence Lessig was the first-but certainly not the last-to argue that the best way to deal with legal questions arising from the Internet is to rely on the common law process and other traditional legal tools.22

This Article uses existing case law to argue in favor of the common law process, in which I include both judge-made common law and judicial interpretation of existing statutes. I suggest in particular that judicial adaptations are likely to be more successful than federal legislative solutions. Three years after Lessig urged that we rely on the common law, Mark Lemley commented that the "hundreds of reported decisions" were "arguably doing a pretty good job of adapting existing law to the new and uncertain circumstances of the Net."23 Three years later, we have thousands of cases and an overlay of several new federal statutes.

Following the path of Lessig and Lemley, I suggest in this Article that judicial responses-including both common law development and the judicial interpretation of existing statutes-to the novel questions posed by the Internet have been adequate to the task, and that the newest statutes are at best unnecessary and at worst shortsighted. These instant legislative solutions offer the promise of speed, but the speed is often expended careening around blind corners. By contrast, the common law method offers the luxury of time and successive experience, while still dealing with the concrete problems posed by the new technology.

This Article illustrates the dangers of over-hasty congressional action by focusing on illustrative federal responses to cyberlaw dilemmas. Parts I and II examine two recent federal enactments that are at best useless and at worst harmful in dealing with the problems that prompted them. Part III offers a cyberspace common law success story: in the absence of clarifying legislation, state and federal courts have been struggling with the question of personal jurisdiction in the context of the Internet, and, after some early missteps, seem to be approaching a rational solution.

I. LOST IN CYBERSPACE: THE ACPA

The Internet brought with it a new form of piracy, quickly labeled "cybersquatting" or "cyber-piracy." Cybersquatting can occur when an individual or company registers as a second-level domain name either a well-known trademark or a typographical

variation on that trademark.24 (A second-level domain name is what immediately precedes the last "dot": for example, "vanderbilt" in the domain name "www.vanderbilt.edu.") The cyber-pirate often then tries to sell the domain name to the trademark owner. Sometimes the cyber-pirate tries instead to divert the trademark owner's would-be customers to his or her own web site in the hopes of making a profit. Yet sometimes the registrant has a legitimate reason for registering the name, which makes defining cybersquatting difficult. Enterprising individuals have registered the names of celebrities ("juliaroberts.com"25), businesses ("panavision.com"26), and organizations ("plannedparenthood.com"27), as well as such variants as "sportillustrated.com"28 and "lucentsucks.com."29

To protect trademark owners from this new activity, Congress passed the Anticybersquatting Consumer Protection Act ("ACPA"), which took effect in November 1999. The ACPA addresses the problem by creating a cause of action by a trademark owner against anyone who "registers, traffics in, or uses" a mark as' a domain name with "a bad faith intent to profit from" the mark. The statute requires that the trademark be either distinctive or famous, and that the domain name be identical, confusingly similar to, or-in the case of a famous mark-dilutive of the mark. Under the ACPA, bad faith is determined by looking to a nonexhaustive list of nine factors that attempt to measure the legitimacy of the registrant's claim to the name and any evidence that the registrant has tried to profit from the name.

The ACPA is a superficially attractive statute. It was enacted quickly with a great deal of popular support, since it is hard

to be in favor of cyber-piracy. But like several other recent public relations enactments by Congress,30 it was also redundant insofar as it dealt with a problem already sufficiently addressed by other statutes, and confusingly drafted because of the pressure of the perceived "crisis." With the possible exception of overcoming occasional jurisdictional problems--an issue I address later31--longstanding trademark infringement law and the relatively new Federal Trademark Dilution Act (FTDA) could, and in many cases did, produce identical results in all cases subject to the ACPA. But Congress did not care that trademark law was developing solutions to cybersquatting; instead, it jumped in prematurely, making cyberlaw now more, rather than less, complicated and confused.

In Part I.A, I examine cases against cybersquatters under both existing law and the new ACPA in order to demonstrate that the ACPA was unnecessary at best. But as might be expected from a statute drafted hastily with great public and industry support, its problems go beyond mere redundancy. In Part LB, I examine the harm that the ACPA is causing to the intangible interests that remained largely unrepresented in Congress's attack on the problem. Finally, beyond these flaws, the ACPA is also chasing technological and other innovations. For that reason, it is no better at dealing with the newest problems than were the prior statutes, and has already been eclipsed by other remedies such as a new dispute resolution procedure. I consider these problems in Part LC. In Part LD, I return to the theme of this Article by examining Congress's attitude, in passing the ACPA, toward the intangible concerns I have identified.

A. The ACP Was Unnecessary

1. Introduction

Perhaps the clearest evidence that courts were perfectly capable of responding to cybersquatting without the ACPA is the case of Sporty's Farm v. Sportsman's Market.32 Sportsman's Market, a catalog business selling aviation equipment, tools, and home accessories, had been using the mark "Sporty's" since the 1960s and had registered it in 1985.33 In 1995, a businessman who received Sportsman's catalogs decided to enter the aviation catalog business himself and began a company that registered the name "sportys.com."34 That company later sold the name to Sporty's Farm, a wholly-owned subsidiary which grows and sells Christmas trees.35 When Sporty's Farm brought a federal declaratory judgment action against Sportsman's Market seeking the right to continue using the name, Sportsman's Market counterclaimed for trademark infringement and trademark dilution.36 The ACPA had not yet been enacted.

The district court ruled in favor of Sportsman's Market on the trademark dilution count, finding that the mark was-as required under the FTDA-both famous and distinctive, and that the registration of sportys.com diluted the mark because it "compromise[d] Sportsman's Market's ability to identify and distinguish its goods on the Internet."37 The court ordered Sporty's Farm to relinquish the domain name, which Sportsman's Market subsequently acquired. But the court refused to grant either damages or attorneys' fees because Sporty's Farm's conduct was not "willful" under the FTDA.

Both parties appealed to the Second Circuit. Judge Calabresi eventually issued a unanimous opinion for a three-judge panel, affirming both of the district court's substantive rulings: Sportsman's Market was entitled to the domain name, but it was not entitled to damages or attorneys' fees. What makes Judge Calabresi's opinion notable is that it applies the newly enacted ACPA rather than the

FTDA.38 Nevertheless, it found nothing to disturb in the trial judge's FTDA ruling. At least in this case, everything that the ACPA accomplished could be done-indeed, had been done-under the FTDA.

Sporty's Farm provides a particularly cogent example of the overlap between the FTDA and the ACPA, but it is not alone. In the six years since the passage of the FTDA, courts have used it repeatedly in cybersquatting cases, by and large reaching intuitively correct results-or at least results that are unlikely to differ markedly from those produced by the ACPA. An examination of how the courts have done so illustrates the strength of the common law method: the courts have applied existing statutes and doctrines to new problems, one case at a time, and in doing so have constructed a jurisprudence that looks remarkably like the congressional enactment without suffering from its flaws.

There are five key differences between the FTDA and the ACPA, but for all but one of these differences the courts have interpreted the FTDA in a way that presaged the ACPA.39 First, trademark dilution and cybersquatting-the targets of the respective statutes-are not identical: dilution as defined by the FTDA is the commercial use of a trademark in a way that harms its reputational value, while cybersquatting as defined by the ACPA is the bad faith registration of a trademark as a domain name. Second, in order to prevail under the FTDA, a trademark owner must show that its mark is both famous and distinctive; the ACPA requires only that it be either famous or distinctive. Third, the ACPA incorporates a standard of "bad faith intent to profit" for any recovery, even injunctive relief, while the FTDA has no state-of-mind prerequisite for injunctive relief but requires "willfulness" before a court can award damages, costs, or attorneys' fees. Fourth, the FTDA provides for ordinary injunctive relief as the standard remedy, while the ACPA authorizes courts to order the transfer of the domain name from the defendant to the plaintiff.40 Finally, the ACPA ex

illicitly allows in rem suits against the domain name itself, but only if the registrant of the name either cannot be located or is not subject to the court's jurisdiction. I deal with each of these differences in turn.

2. Commercial Dilution Versus Bad Faith Registration

The poster child for the crusade against cybersquatting is Dennis Toeppen. An enterprising netizen, Toeppen was an early registrant of over 100 domain names, including some incorporating the trademarks of such companies as Delta Airlines, Neiman Marcus, and Eddie Bauer. But Toeppen's lasting claim to fame arose from his registration of the names "panavision.com" and "panaflex.com." His web page for panavision.com displayed photographs of the city of Pana, Illinois; his web page for panaflex.com displayed the word "Hello." Not surprisingly, Panavision (which owns registered trademarks for both Panavision and Panaflex) sued Toeppen under the FTDA in 1996. Both the district court and the Ninth Circuit Court of Appeals recognized the novelty of Panavi

sion's claim under the FTDA, but neither court had any difficulty ruling in favor of Panavision.

The most difficult question for the Court of Appeals was whether Toeppen's use of the trademark was "commercial" within the meaning of the FTDA.41 Some earlier precedent had held that mere registration of a domain name was not a commercial use under the act.42 But the court reasoned that Toeppen was using the name commercially: although he was not using it to sell products, he was hoping to make money from it by selling the name back to Panavision. The ACPA incorporates this insight by making both attempts to sell the name to the mark owner and registration of multiple trademarked domain names relevant to the determination of bad faith. The Panavision court did not need the assistance of the ACPA, however, to perceive that Toeppen was making a commercial use of the name-nor have numerous other courts before and since.43 One need not even offer to sell the domain name to come within the ambit of the FTDA: in Planned Parenthood Federation of America v. Bucci,44 the court held that an anti-abortion web site

owner who used the domain name "plannedparenthood.com" was using the mark commercially because, among other things, his "actions are designed to, and do, harm plaintiff commercially."45 Other courts have found trademark infringement in similar circumstances. One court has held that trademark infringement "does not require that Defendant actually caused goods or services to be placed into the stream of commerce," because the Lanham Act encompasses "the dissemination of information, including purely ideological information."46 Thus neither the FTDA requirement that the mark be used commercially, nor the requirement under infringement law that the mark be used commercially and "in connection with the sale or offering for sale, distribution, or advertising of goods and services,"47 serves as an obstacle to finding liability in cybersquatting cases.

The Ninth Circuit also had to decide whether Toeppen's use of the trademark as a domain name constituted dilution. Again, the court had little trouble concluding that Toeppen's conduct diluted the Panavision mark even though it did not fit the traditional definitions of tarnishment or blurring.48 The existence of Toeppen's web sites both "put[ ] Panavision's name and reputation at his mercy,"49 and diminished Panavision's ability to use its mark to lead customers to its site. Customers looking for Panavision might have to wade through the hundreds of web sites likely to be produced by a search engine, or might become frustrated with Panavision-and discouraged from further searching-if panavision.com did not lead them to Panavision's home page. In short, Toeppen's use of the Panavi

sion marks as domain names lessened Panavision's capacity to identify itself and its products on the web. Other courts have reached similar conclusions under the FTDA.50 One court has held that "internet cyberpiracy constitutes per se trademark dilution."51 In addition, many courts have found cybersquatting to be trademark infringement as well as (or instead of) trademark dilution.52 Indeed, in one case a court found that the use of the plaintiffs mark ("Seawind" for a kit-built amphibious airplane) was not infringed by the defendant's physical products (a "Seawind Builders Newsletter" and a plane called a "Turbine Seawind," built from plaintiffs kit with the addition of a turbine engine), but was infringed by the mere use of the plaintiffs mark in defendant's domain name ("seawind.net").53

As the Panavision case illustrates, a dilution action can be strong medicine. Indeed, a number of commentators have criticized the FTDA for overprotecting trademarks from noncompeting uses.54

At least in the context of domain name disputes, however, the courts have been quite careful to limit the definition of dilution to cybersquatting and related acts. Another Ninth Circuit case, Avery Dennison Corp. v. Sumpton,55 provides an illuminating contrast to Panavision. The plaintiff sold office products under the trademarked names "Avery" and "Dennison," and maintained web sites at "avery.com" and "averydennison.com." The defendant, Sumpton, was a provider of vanity e-mail addresses, leasing e-mail addresses to individuals that incorporated their last names, their careers, their hobbies, and so on.56 He had registered thousands of domain names for this purpose, including both "avery.net" and "dennison.net." He did so because both Avery and Dennison are relatively common last names. He never tried to sell either name to Avery Dennison. Although the district court had granted summary judgment for Avery Dennison, the Ninth Circuit reversed and ordered the lower court to grant summary judgment to Sumpton. The Court of Appeals held that Sumpton was not using the trademarks commercially: although he was making money from his registration of both domain names, he was not "capitalizing on [their] trademark status."57 Thus, although what Sumpton did would ordinarily be considered "commercial," the court limited the FTDA to protect legitimate uses of words that happened to be trademarked.

In Avery Dennison Corp. and Panavision, then, the court struck a careful balance: preventing cybersquatting but allowing

registration of trademarked domain names for other legitimate reasons. In Avery Dennison Corp., the court interpreted "commercial" to strike this balance. As we will see in the next section of this Article, courts have also relied on other aspects of the FTDA to strike at cybersquatters without endangering legitimate domain name registrations.58

3. Famousness and Distinctiveness

To prevail on a claim of trademark dilution under the FTDA, the plaintiff must show that its mark is famous and that the defendant's use dilutes the distinctiveness of the mark.59 The FTDA provides that fame and distinctiveness may be determined by considering a nonexhaustive list of eight factors.60 Unlike trademark infringement, which applies only to confusing uses of the mark, the FTDA prevents all commercial uses of the mark. Thus, only the use of "Nike" on shoes is likely to be trademark infringement, but Nike sodas, cars, and television sets all clearly violate the FTDA. Because of its breadth, the FTDA is meant to apply only to truly famous marks in order to preserve the supply of available words. Both the legislative history and many cases interpreting the FTDA support this narrow definition of famousness.61

As an abstract matter, then, the FTDA might seem to be unavailable in many cases of cybersquatting.62 While "nike.com" and 44 panavision.com" might satisfy the famousness component of the FTDA, names that are recognized only within an industry or within a local geographic area might not. Nevertheless, in the context of domain name disputes, courts have frequently found such local or industry-limited marks to be famous for purposes of the FTDA. Examples of such limited-fame marks include "Teletech" for "the largest provider of primarily inbound integrated telephone and Internet customer care worldwide"63 and "La Opinion" for "the largest circulation stand-alone Spanish language newspaper in the United States."64

The danger, of course, is that in an attempt to reach cybersquatting, courts might construe famousness so broadly as to undermine the intent to limit the applicability of the FTDA.65 However, most courts have extended the FTDA to limited-fame marks only where the junior user (including a domain name registrant) has no plausible claim to the use of the mark-that is, where the junior user's most likely motivation was to benefit from the public recognition of the mark. There is nothing in the FTDA that explicitly requires this result, and evaluating the junior user's claim to the mark-since it does not fall among the FTDA's eight listed factors-would not seem particularly relevant to a determination of famousness.66 Nevertheless, an examination of actual FTDA cases

strongly suggests that a court's determination of famousness, for less than universally famous marks like Nike, is influenced by the court's view of the legitimacy of the junior user's motives.

For example, in Star Markets, Ltd. v. Texaco, Inc., a court found that "Star Markets" was not sufficiently famous to satisfy the FTDA.67 The name apparently had considerable name recognition and famousness in Hawaii, and probably would have met the standards applied in the Teletech and La Opinion cases. But Texaco had slowly been converting its mainland convenience stores (associated with its gasoline stations) to "Star Mart" in order to resonate with its "Star-T" logo.68 Texaco thus had a strong claim that it was not trading on Star Markets' name recognition or goodwill, and the court rejected Star Markets' FTDA claim. On the other hand, in WAWA Inc. v. Haaf, the court found (with no discussion) that the use of WAWA for convenience stores in a five-state area was sufficiently famous to satisfy the FTDA, and thus enjoined defendant Haaf's use of the name HAHA for a competing convenience store.69 Although Haaf claimed that he called his market "HAHA" because he and his wife intended to own it jointly, no partnership or other documents had been prepared, and he was the sole owner of the store and the sole registrant of the fictitious name. It is easy to conclude that Haas use of a name similar to WAWA had no legitimate purpose, but instead was deliberately designed to benefit from WAWA's mark.

Courts have reached similarly divergent results under the FTDA as applied to domain name disputes. In Avery Dennison Corp., for example-in which, as discussed earlier, the Ninth Circuit found that Sumpton's registration of avery.net and dennison.net for vanity e-mail addresses was not a commercial use-the court also held that neither "Avery" nor "Dennison" was sufficiently famous to warrant FTDA protection.70 But another court found "NYSSCPA" (for New York State Society of Certified Public Accountants) famous in the context of its use on a web site by an ac

countant placement firm seeking to capitalize on name recognition within the niche market of accountants in New York.71

While no court has explicitly held that the junior user's motivation is relevant to the question of famousness, a fair reading of the cases suggests that such an analysis explains the pattern of different holdings in factually similar cases.72 Moreover, several courts have noted that fame in a "niche" market is more likely to support an FTDA claim in cases in which the two users of the mark are direct competitors.73 Distinguishing between uses depending on the competitive status of the user supports the notion that the FTDA is being interpreted to take into account the junior user's motives: a competitor who uses a protected mark (other than a truly famous mark such as Nike) is more likely to be trying to gain a competitive advantage than is a noncompetitor who uses the mark.

And except for its relevance in revealing the defendant's motives in using the mark, the fact that the plaintiff and defendant are competitors is actually inconsistent with the purposes of the FTDA, as several courts have noted.74 The FTDA was designed to prevent the appearance of Dupont shoes, Buick aspirin, and Kodak pianos. Such uses would not involve consumer confusion, and therefore would not fall under traditional trademark infringement analysis. The statute was not meant "as mere fallback protection for trademark owners unable to prove trademark infringement"75 because of a lack of sufficient confusion, but rather as a remedy for

a specific harm outside the purview of infringement. This narrow purpose belies any interpretation that would make relief more available to competitors than to noncompetitors.76 Thus, to the extent that the owner of a marginally famous mark is more likely to succeed in an FTDA suit against a direct competitor than against a noncompetitor, the cases can best be explained by positing that a direct competitor is more likely than a noncompetitor to have had nefarious motives for choosing the mark.

This accommodation between the extent of the mark's fame and the legitimacy of the junior user's claim presages the requirements of the ACPA. While the ACPA provides protection to marks that are either famous or distinctive, its primary focus is on the motives of the domain name registrant: it punishes only a bad faith intent to profit. Had Star Markets and WAWA involved domain name disputes under the ACPA, it is likely that courts would have reached exactly the same results as under the FTDA. The registration of "starmart.com," for example, raises an issue that neither the FTDA nor the ACPA were designed to remedy: multiple, legitimate uses of the same mark are possible in the real world (United Airlines and United Van Lines, for example), but not in the virtual world (only one company can register "united.com"). When confronted with this problem, under either the FTDA or the ACPA, courts have consistently allowed registration of a multiple-use mark on a first-come, first-served basis.77 Again, then, the ACPA has not added anything to judicially developed FTDA doctrines.

If the ACPA was not necessary to protect trademark owners, neither has it changed many results. Courts interpreting the ACPA have sometimes carried over from the FTDA both the weak definition of fame in cybersquatting circumstances and the concern about

the potential breadth of dilution causes of action. Although the ACPA decreases the importance of a finding of fame since the cause of action lies to protect trademarks that are either famous or distinctive-courts have nevertheless generally found marks to be both famous and distinctive even in situations in which "fame" is relatively modest: for example, the use of "Joe Cartoon" for cartoon graphics.78 Thus, even when applying the ACPA, courts have made findings that would justify relief under the FTDA. Conversely, like the FTDA, the ACPA limits the supply of available words by protecting even noncompeting uses of trademarked domain names. The requirement that the cyber-pirate be shown to act in bad faith affords some limit on the breadth of the cause of action. But at least one court has also suggested that "famousness" under the ACPA should be narrowly interpreted in order to avoid "granting `rights in gross' in a trademark."79 Thus, even if Congress's intent was to grant more protection under the ACPA than under the FTDA, courts have not generally done so-and with good reason, since extending protection for nonconfusing uses of marginally famous marks is most easily justified when the junior user is attempting to appropriate for itself the commercial advantages of the mark.

As with the requirements of commercial use and a showing of dilution, then, the FTDA requirement that a mark be both famous and distinctive has not deterred courts from giving relief in cybersquatting cases. Indeed, I have not been able to locate a domain name dispute case in which a defendant found liable under the ACPA would have escaped liability under the prevailing interpretation of the FTDA because of a lack of fame, commercial use, or actual or potential dilution.80

4. Intent Requirements

The defendant's intent in using the trademark plays a different role in infringement, dilution, and ACPA cases. For infringement, the alleged infringer's intent to copy the mark and thereby create confusion between its product and the plaintiffs is relevant to the question of whether there is a likelihood of confusion.81 Under the FTDA, the plaintiff must demonstrate the defendant's "willfulness" only if plaintiff seeks damages, costs, or attorneys' fees; the defendant's intent is statutorily irrelevant to injunctive relief.82 Finally, the ACPA imposes liability only for a "bad faith intent to profit" from the use of the mark; good faith uses are not actionable.83

The ACPA requirement of bad faith is designed to narrow the breadth of the statute and to avoid penalizing anyone except classic cyber squatters.84 In order to avoid giving greater protection to trademark owners, then, the ACPA interposes an additional prerequisite to recovery. For that reason, the ACPA's different intent requirement is unlikely to offer any greater protection to trademark owners than is provided by the already-existing FTDA and Lanham Act. Imagine that Toeppen, our cybersquatting entrepreneur, had-after a dispute with Panavision about a defective television set-simply registered panavision.com and connected that domain name to scenic views of Pana. In this counterfactual scenario, he did not try to sell the name back to Panavision, nor did he register any other trademarked names. Under the ACPA, it is unclear whether Panavision could prevail, since it would be difficult to demonstrate Toeppen's bad faith intent to profit. Of the nine factors listed in the ACPA for determining bad faith, it seems likely that four would favor Toeppen and five would favor Panavision. Nevertheless, the mere existence of the web site at that URL address

would dilute Panavision's trademark. It is thus a clearer violation of the FTDA than of the ACPA.85

Nor does the ACPA offer greater protection for innocent domain name registrants, since a plaintiff can choose to file suit under either statute. That our hypothetical Toeppen would probably prevail on an ACPA claim does not preempt Panavision from suing him under the FTDA instead.

Thus, since the ACPA neither provides greater protection to mark owners nor reduces the protection available under the FTDA, the ACPA seems paradigmatically redundant, at least as far as the different intent requirements are concerned. Moreover, as noted earlier, courts have essentially created an intent requirement under the FTDA unless the disputed trademark is unequivocally and universally famous. Thus, the different intent requirements do not end up making much difference to the results.

Whenever a statute sets out an intent test, moreover, it cannot help but create uncertainty at the margins. Cello Holdings v. Lawrence-Dahl Cos.86 is a typical example. Plaintiff, a high-end audio equipment provider, had registered "Cello" as a trademark. The defendant sold vintage audio equipment online under the name Audio Online, and also had a sideline business: he registered numerous domain names-mostly those not trademarked-with NSI, including "gotmilk.com," "4nasdaq.com," "thenyse.com," and "goldmetal.com," and offered them for general sale on the Internet. This second part of his business had resulted in the sale of two domain names for a total of $1,100. In 1997, defendant tried to register the names of at least twenty musical instruments, including "violin.com" and "drums.com." Only "cello.com" was available, however, and defendant obtained it. He then offered cello.com for sale to at least nine companies or individuals, including the plaintiff, for about $5,000. Plaintiff replied to this solicitation with a federal lawsuit under the FTDA and the ACPA.

In denying Cello Holdings' motion for summary judgment, the court held that numerous genuine issues of fact existed, including both the fame and distinctiveness of the cello mark. It also found a genuine dispute of fact with regard to bad faith-and with good reason. It is hard to tell whether this defendant was an enterprising capitalist, purchasing apparently unowned goods that others might want and then looking for buyers, or a cyber-pirate who stole trademarks from their rightful owners in hope of holding them for ransom. Most cybersquatters seem to ask for tens-if not hundreds-of thousands of dollars,87 and most target potential buyers by zeroing in on particular trademarks. Perhaps this defendant was simply an inept cybersquatter, but it is also possible that this is not the type of activity that Congress meant to proscribe.

In any case, the different intent requirements do not undermine my contention that the ACPA was unnecessary because it duplicates remedies available under other statutes. At best, the ACPA seems to offer no more assistance to an aggrieved trademark holder; at worst, some plaintiffs may have a paradoxically harder time using the ACPA because of its heightened intent requirement.

5. Remedies

The ACPA also provides remedies that differ from those provided under the FTDA. The remedy for ordinary trademark dilution is normally an injunction against further use of the mark by the defendant.88 In cybersquatting cases, however, the plaintiff usually also wants to be able to use its own mark as a domain name. Again, courts have had no difficulty under the FTDA ordering defendants to relinquish all rights to the disputed domain name, thus freeing it

for subsequent use by the plaintiff.89 Courts have also ordered defendants to transfer domain names to plaintiffs.90 And in at least one FTDA case, Network Solutions, Inc. ("NSI"), the private company in charge of registering domain names at that time, had "deposited complete control and authority over the disposition and use of the [disputed] domain name" into the court itself91--an action specifically contemplated by the not-yet-enacted ACPA but not by the FTDA.

6. In Rem Jurisdiction

Suits against cybersquatters present potentially difficult questions of personal jurisdiction. While most state long-arm statutes are broad enough to reach conduct, such as trademark infringement or dilution, which has an effect inside the states, the constitutional requirements under the Due Process Clause are somewhat more limiting. Black letter law requires that in order to be subject to personal jurisdiction, the defendant must have minimum contacts with the forum state such that the court's exercise of jurisdiction is consistent with traditional notions of fair play and substantial justice.92 As I discuss in Part III, courts initially struggled with questions about the extent to which the creation of a web site constitutes minimum contacts with any forum in which it is accessible. As courts reached a consensus that the establishment of a "passive" (or noninteractive) web site is not by itself sufficient to confer jurisdiction, the question of jurisdiction over cybersquatters became more complex: the classic cybersquatter often does no more

than register the domain name or, as Dennis Toeppen did, create a passive and noncommercial web site.

To avoid this potential problem with jurisdiction, the ACPA provides an alternative to an in personam action against the domain name registrant: it creates an in rem action against the domain name itself. The in rem action is available only if the plaintiff, after due diligence, either cannot locate the domain name registrant or "is not able to obtain in personam jurisdiction over" the registrant.93 The statute does not make clear whether the plaintiff must show only that the court in which it has filed lacks personal jurisdiction or whether the plaintiff must show that no American court can exercise jurisdiction over the defendant. The only court that has directly addressed the issue has concluded that the in rem action is available only if no American court has jurisdiction.94 For this reason, the in rem provision is likely to apply in only a limited number of cases, that is, where the registrant is unknown or foreign.

In fact, courts have had little trouble exercising personal jurisdiction over cybersquatters even absent the ACPA's in rem provision. Courts have consistently held-in domain name disputes as well as other cases-that the mere registration of a domain name, or the establishment of a "passive" web site at that URL, does not constitute minimum contacts with any other state, including the state in which the registration occurred.95 Courts have required that the defendant do something more, somehow directing its actions toward the forum state. But courts confronted with bad-faith registration of domain names-cybersquatting-quickly found ways to conclude that the defendant had targeted the forum state.

An offer to sell the domain name back to the trademark owner, for example, has been held to constitute activity directed toward the trademark owner's home state.96 Even in the absence of attempted extortion, cyber-pirates are usually subject to personal jurisdiction in the trademark owner's state. Following a seminal Seventh Circuit decision in a non-Internet context, Indianapolis

Colts, Inc. u. Metropolitan Baltimore Football Club Ltd. Partnership,97 a number of courts have found that the act of registering a domain name that intentionally targets a particular trademark owner constitutes an act directed toward the trademark owner's home state.98

Courts have also found jurisdiction over foreign registrants if the foreign registrant uses the domain name to enter the U.S. market. In Quokka Sports, Inc. v. Cup International, Ltd., for example, the court held that it had personal jurisdiction over a New Zealand company.99 The company's web site included advertisements for at least ten American companies (including several that did business only in the United States), and when it negotiated for advertisers, the company quoted prices in American dollars.100 The court concluded that the company, by targeting American consumers, had purposefully availed itself of the benefits of the United States commercial forum.101

As with the attenuation of the "famousness" requirement, there is a danger that a desire to reach cyber-pirates might distort doctrines of personal jurisdiction. But courts have been careful, under both the ACPA and its predecessors, to distinguish between cyber-pirates and more innocent trademark infringers. As noted earlier, courts dealing with cybersquatters have found that registra

tion of a particular trademark as a domain name intentionally targets the trademark owner, thus constituting a sufficient additional contact with the forum to create personal jurisdiction. In the absence of evidence of bad faith, however, courts have been unwilling to extend this principle to all those who simply register another's trademark as a domain name. If the domain name registrant is unaware of the trademark or has a legitimate reason to choose the mark as a domain name, courts consistently require more traditional indicia of contacts with the forum state than simple registration of a trademarked domain name.102

Here again, the ACPA has not altered doctrines of personal jurisdiction: in personam suits, at least, are neither easier nor harder to bring under the ACPA than under previous trademark statutes.

There remains, however, the question of domain name registrants who cannot be located, or foreign registrants with virtually no other contacts in the United States. Although plaintiffs have successfully used the in rem provisions against foreign registrants, in many cases the same result could have been reached under the developing doctrines of personal jurisdiction in cybersquatter cases. In BroadBridge Media v. hypercd.com, for example, the Canadian defendant offered to sell the domain name to the plaintiff, engaging in negotiations over several days.103 The Ninth Circuit in Panavision held similar conduct sufficient to confer jurisdiction. In Heathmount A.E. Corp. v. technodome.com, the plaintiff alleged-as required under the ACPA-that the defendant had registered its domain names "in a bad faith attempt to profit from use of Plaintiffs marks."104 Other courts, taking the plaintiffs allegations as

true, have found such an intent to be a sufficient contact with the plaintiffs home state to satisfy due process under the Indianapolis Colts theory. Thus, although in both BroadBridge Media and Heathmount A.E. Corp. the courts held an in rem suit appropriate because personal jurisdiction was lacking, it is probable that in the absence of the ACPA the courts would have examined the precedent more carefully and found the requisite personal jurisdiction over the defendants.105

Only if the plaintiff fails to allege bad faith-or if there is sufficient evidence to overcome that allegation-does the problem of personal jurisdiction become acute. But courts have generally held that an allegation of bad faith is necessary even in an in rem suit, dismissing such suits where bad faith is absent.106 Moreover, the purpose of the ACPA is to remedy cyber-piracy, not innocent trademark infringement, and thus the in rem provisions should not be invoked in the absence of bad faith.

What about mark owners who cannot locate the registrants of infringing domain names? I have found only two cases in which plaintiffs attempted to use the in rem provisions of the ACPA in that circumstance.107 In one case, the domain name registrant never responded, and the plaintiff was awarded the name in a default judgment.108 In the other, the court dismissed the action, finding that the plaintiff had not exercised due diligence in searching for the registrant.109 The plaintiff later found the registrant and sued him in personam.110

Finally, had Congress not enacted the ACPA, it is at least possible that courts would have developed a doctrine of in rem jurisdiction on their own. At least two plaintiffs' lawyers' creativity

led them to bring in rem suits against domain names, alleging various trademark violations, prior to the enactment of the ACPA. In one case, the court allowed the plaintiff to amend its complaint some nine months later to allege a claim under the newly enacted ACPA, thus avoiding the need to decide whether the original in rem action could be maintained.111 In the other case, the district court dismissed the action, holding that the FTDA did not permit suits in rem.112 Plaintiffs appealed, but the intervening enactment of the ACPA spared the court of appeals from reaching the question: it remanded the case for consideration of the application of the ACPA.113 In the absence of the ACPA, perhaps courts would ultimately have concluded that in rem actions were available for trademark infringement or dilution by domain names.114

The in rem provisions of the ACPA, then, unlike the substantive provisions, do add to the previous remedies afforded trademark owners, but not significantly. The provisions make it easier to sue foreign defendants, although those defendants would probably be amenable to jurisdiction under existing doctrines. The provisions may also prove useful in the case of unknown or unlocated domain name registrants, but the record to date suggests that those cases may be rare.

In light of the additional utility-however limited--of the in rem provisions, I would have no quarrel with them if they were well-drafted to avoid confusion. As I discuss in the next section, however, the in rem provisions are causing confusion and complexity. That confusion is too high a price to pay for a statute that adds so little to existing law.

B. The ACPA Is Causing Harm

Most statutes have advantages and disadvantages. Under ordinary circumstances, we expect that Congress will pass a law only if its perceived advantages outweigh its perceived disadvantages. In regulating cyberlaw, however, where the disadvantages are largely intangible, we might expect laws to slip through that, on closer examination, do not satisfy this cost-benefit analysis. The ACPA is such a statute.

I have just argued that the advantages of the ACPA are minimal. Its disadvantages are not overwhelming, either, but they are greater than its almost nonexistent benefits. The ACPA is causing two kinds of harm, both difficult to identify or measure if one looks only to tangible interests such as those of trademark owners, consumers, and cybersquatters.

First is the harm to the legal regime itself. Whenever Congress enacts a new statute affecting actions that are already subject to other laws, some preliminary confusion is bound to occur. When faced with the new statute, courts must interpret the new language and figure out how the new statute fits into, adds to, or modifies the existing regulatory regime. The ACPA is no exception.

The most significant source of confusion is the in rem provisions. Courts have split first on where the ACPA allows an in rem action to be brought. Several courts have focused on the language of section (2)(A), which provides:

The owner of a mark may file an in rem civil action against a domain name in the judicial district in which the domain name registrar, domain name registry, or other domain name authority that registered or assigned the domain name is located .... 115

These courts have held that in rem jurisdiction lies only in the district in which NSI or its successors reside.116 At least one court has

expressed doubts about the constitutionality of any broader jurisdictional interpretation: after an extended discussion of Shaffer v. Heitner,117 the court concluded that the minimum contacts test of International Shoe118 applied to in rem as well as quasi in rem actions.119 Whether or not that is a fair reading of Shaffer,120 the injection of such a difficult question into the domain name cases undermines the utility of the ACPA in rem proceeding as a streamlined method of resolving domain name disputes. The conclusion that suits can only be brought in the registrar's home district, moreover, may overburden a few federal courts. In any case, there is a division of authority: other courts, without discussion, have allowed in rem suits in districts other than the registrar's home district, as long as the court obtains control over the domain name.121

Another question also arises under the in rem provisions. By its terms, the ACPA only permits in rem suits if personal jurisdiction cannot be obtained over the domain name owner (or if the owner cannot be found).122 Courts have therefore confronted the

question of the burden of proof required to find an absence of personal jurisdiction. The two courts that have addressed the issue have held that the plaintiff bears the burden of proving, by a preponderance of the evidence, that it is not able to obtain in personam jurisdiction over the defendant.123 But since the plaintiff traditionally also bears the burden of proving the existence of jurisdiction,124 that potentially leaves plaintiffs in a rather difficult situation.

The allocation of the burden of proof to one party can mean that in close cases that party will lose. So if the question of personal jurisdiction over the defendant is close, a plaintiff may find itself without any cause of action under the ACPA: unable to definitively prove or disprove personal jurisdiction, it will not be able to bring either an in personam or an in rem action. While this Catch-22 situation may seem only a remote possibility, it does undermine the usefulness of the ACPA to potential plaintiffs seeking the promised easy remedy for cybersquatting, as well as illustrate the problems that may arise from hastily drafted statutes.

The jurisdictional difficulties in close cases are compounded by another ruling by one of the two courts that have addressed the burden of proof question. Confronted with an attempt to bring in rem and in personam suits simultaneously, the court easily concluded that because the ACPA clearly provides for in rem jurisdiction only if in personam jurisdiction is lacking, the two alternatives are "mutually exclusive avenues for relief against putative infringers."125 That much seems incontrovertible. But the court went on to hold:

In this regard, a mark owner may not simultaneously file an in rem cause of action and an in personam claim in the hope that one claim will survive the court's jurisdictional inquiry. Rather, a mark owner must choose prior to filing whether to proceed in rem against the domain name or in personam against a putative infringer.126

Thus, a plaintiff not only bears the burden of proving or disproving jurisdiction over the defendant; the plaintiff also has to guess in advance which way the court will rule.127

This type of confusion is the inevitable consequence of a new statute, a confusion only exacerbated by poor drafting. If my thesis is correct, the courts will eventually sort it out and reach a reasonable and predictable interpretation of the in rem provisions of the ACPA. But even if only temporary, conflicting, confusing, or unpredictable court decisions do some damage to the stability of the legal landscape. Such detrimental effects, however minor, should be justified by a corresponding benefit from the statute. I suggested in Part LA that the ACPA offers no commensurate benefit.

Moreover, the ACPA is also causing more substantial harm to a more important intangible interest. What is particularly troubling about some of the confusion generated by the ACPA is that it implicates the First Amendment. A prohibition against paradigmatic cybersquatting-the intentional registration of another's trademark in order to profit commercially-presents no more First Amendment problems than would use of the trademark on a competing good in order to deceive the public. But sometimes a trademark is used within a domain name not for commercial purposes but to make a statement about the trademark holder. Examples include the registration of "plannedparenthood.com" by an antiabortion activist; the registration of "lucentsucks.com" by a critic of Lucent Technologies; and the registration of "morrisonandfoerster.com" (and related names) by an opponent of "corporate America" and its lawyers. The web sites identified by such domain names are sometimes called "gripe sites."

Prior to the ACPA, the courts had a mixed record of sensitivity to free speech issues raised by gripe sites. Under traditional trademark law, use of a trademark as "part of a communicative message" would be protected.128 In non-Internet trademark cases, a

balancing test ensured that the public interest in free expression could rarely be overcome by legitimate parody or political criticism.129 But even before the ACPA took aim at domain name disputes, courts were viewing Internet political expression with a much more skeptical eye, finding both "commercial use" and likelihood of confusion more easily in cases involving web sites than in non-Internet uses of trademarks.

One early commentator130 tellingly compared Planned Parenthood Federation of America v. Bucci131 with International Association of Machinists v. Winship Green Nursing Center.132 In Planned Parenthood, the court concluded that the public was likely to be confused, even if only temporarily, when entering "plannedparenthood.com" brought to the screen an anti-abortion web site. In Winship Green Nursing Center, by contrast, the court held that there was no likelihood of confusion when management circulated letters on union letterhead implying that employees would be fired, and their union dues raised, if the union won the election.133 After Planned Parenthood, other courts using pre-ACPA trademark law also curtailed web sites that were arguably political criticism and that would likely be protected in any medium other than the Internet.134

The insensitivity to First Amendment interests exhibited by these courts may be a symptom of the newness of the technology. A few cases have recognized a distinction between cyber-piracy and gripe sites.135 Given more time, and perhaps some helpful critical commentary, courts might have reached some accommodation between trademark rights and free speech rights under the FTDA.

For example, it is possible to distinguish Planned Parenthood from another case enjoining an arguably political web site,

Jews for Jesus v. Brodsky.136 In Planned Parenthood, the URL plannedparenthood.com called up a web site that purported to be the home page of the real Planned Parenthood organization, prolonging any potential confusion.137 It was only after scrolling or clicking through several screens that a web-surfer might realize that she had not in fact reached Planned Parenthood. In Jews for Jesus, by contrast, the initial web site called up by the domainname "jewsforjesus.org" clearly proclaimed that it was neither affiliated with, nor in agreement with the tenets of, the plaintiff organization. One might argue that under these circumstances, enjoining "plannedparenthood.com" was more justifiable than enjoining "jewsforjesus.com." Perhaps courts might have eventually settled on this or some other plausible distinction in applying the FTDA and the Lanham Act to domain name disputes. Perhaps they still will.138

In the meantime, however, the ACPA has prolonged the process of accommodation and exacerbated the difficulties. Even to the extent that the ACPA simply duplicates previously available remedies, it nevertheless requires courts to analyze the tension between trademark law and the First Amendment under new statutory language, jeopardizing whatever progress had been made under earlier statutes. To whatever degree courts were beginning to weigh First Amendment concerns against trademark rights in FTDA cases, confronting the new language of the ACPA means that they will have to begin anew.139 In addition, the ACPA gives trade

mark owners another weapon with which to threaten those who establish gripe sites. Verizon, for example, after carefully registering verizonsucks.com, was dismayed when an online magazine registered verizonreallysucks.com. It sent the publisher a letter warning that the magazine's registration of the domain name "infringes Bell Atlantic's valuable trademark rights in the Terizon' mark and violates the new Anticyber squatting Consumer Protection Act."140

Moreover, the ACPA might make it even easier to attack gripe sites. One commentator concludes that "the ACPA has granted trademark owners more control over the use of the marks as domain names than they have for any other use."141 Since courts under prior trademark law were, as we have just seen, already according trademarks somewhat more protection on the Internet than elsewhere, any additional protection afforded by the ACPA will cut further into the domain of freedom of speech.

Some courts have already exhibited confusion about the relationship between bad faith under the ACPA and willfulness under the FTDA,142 essentially reducing the ACPA requirement to "bad faith" rather than "bad faith intent to profit." Since virtually by

definition gripe sites involve bad faith in the broadest sense-at least insofar as they disparage or attack a trademark owner-the ACPA might cover them all. Indeed, one court has held that an admitted intent to "get even with" or "mess with" a trademark owner offers "[t]he most persuasive reason for concluding that [the defendant] acted with bad faith intent."143 The ACPA's potential incursion into protected speech is thus a troubling possibility.

To the extent that the ACPA does alter the preexisting law, then, it does so in ways that either increase confusion or heighten the tension between trademark protection and the First Amendment. Again, this seems to be a high price for such a small gain.

C. The ACPA Is Already Becoming Obsolete

1. Cyber-Mischief Innovations

Congress may act more quickly than the courts, but the Internet changes even faster. The final problem with the ACPA is that it deals only with yesterday's problems, not today's. It has nothing to say, for example, about the use of trademarks in metatags. Metatags are computer-readable material, embedded into web sites but unseen by the consumer. They are chosen by the web site owner in order to alert search engines to the contents of the site.

For many people, search engines such as AltaVista or Google provide an entry point into the worldwide web.144 These search engines work by essentially crawling around the web and looking not only at the web site itself, but also at the metatags and other embedded material. (Some search engines look only at the embedded portion, others only at the visible portion, and still others examine both parts.) The search engine stores the metatags it has discovered (and their Internet locations), and when an individual enters words into a search engine it produces a matching list of sites.145

Cyber-pirates quickly realized that putting a competitor's name into their metatags might bring them additional customers. A consumer looking for, say, McDonald's, might type "McDonald's" into a search engine. If Burger King has included "McDonald's" among its metatags, the search engine will produce a list that contains both the McDonald's and Burger King sites.146 A consumer who was just looking for hamburgers-and simply remembered McDonald's as the best-known brand-might decide to buy from Burger King instead. Or perhaps KFC, which doesn't sell hamburgers, might include "McDonald's" and "hamburgers" in its metatags, hoping to remind consumers seeking fast food that there are alternatives to hamburgers. A vegetarian organization might use similar metatags hoping to win converts to its anti-meat positions (imagine the picture of an abattoir that might greet the consumer at a vegetarian web page identified by a McDonald's metatag). Metatags thus offer boundless opportunities for clever marketers.147

Since the public never sees the metatags, however, and the competing web sites are clearly marked with their own trademarks, the use of trademarks in metatags raises difficult questions under infringement and dilution law. How can an unseen trademark cause confusion or dilute the distinctiveness of the mark?148 But confronted with suits alleging that using another's trademark in meta

tags constituted infringement or dilution, the courts have turned to a trademark doctrine known as "initial interest confusion."

The doctrine of "initial interest confusion" was first developed by the Second Circuit in 1987, long before the commercialization of the Internet and the subsequent plethora of Internet trademark issues. In Mobil Oil Corp. v. Pegasus Petroleum Corp.,149 the court held that misleading potential customers through the use of a registered mark could constitute trademark infringement even if any confusion dissipated before a sale was made. Other courts have also adopted the doctrine, allowing a finding of infringement even in the absence of an actual sale.150 It is the use of the trademark to draw attention to one's own products-to get one's foot in the door-that constitutes the infringing use, regardless of what follows.

Despite its development in a non-Internet context, the doctrine of "initial interest confusion" seems tailor-made for metatag disputes. The Ninth Circuit was the first court to draw the analogy between real space and cyberspace:

Using another's trademark in one's metatags is much like posting a sign with another's trademark in front of one's store. Suppose West Coast's competitor (let's call it "Blockbuster") puts up a billboard on a highway reading-"West Coast Video: 2 miles ahead at Exit 7"-where West Coast is really located at Exit 8 but Blockbuster is located at Exit 7. Customers looking for West Coast's store will pull off at Exit 7 and drive around looking for it. Unable to locate West Coast, but seeing the Blockbuster store right by the highway entrance, they may simply rent there. Even consumers who prefer West Coast may find it not worth the trouble to continue searching for West Coast since there is a Blockbuster right there. Customers are not confused in the narrow sense: they are fully aware that they are purchasing from Blockbuster and they have no reason to believe that Blockbuster is related to, or in any way sponsored by, West Coast. Nevertheless, the fact that there is only initial consumer confusion does not alter the fact that Blockbuster would be misappropriating West Coast's acquired goodwill.151

Other courts have also held that the use of trademarks in metatags constitutes infringement in these circumstances.152

There are still open questions, of course, including disputes about the extent to which "initial interest confusion" should apply to sites criticizing the trademark owner.lss And there is a lively debate in the literature about the extent to which metatagging should be considered a Lanham Act violation at all.lsa But despite the superficial strangeness of the metatag question, courts are successfully adapting existing trademark law to this new context. This is another real world problem resolved in the courts, although the ACPA was not prescient enough to add anything to the analysis.

Moreover, neither the ACPA nor a hypothetical Anti-Metatag Act55 are likely to answer the cyberspace trademark questions that are inevitably just around the corner. The abuse of metatags is not limited to trademark infringement: some web site owners load up their metatags with irrelevant but frequently requested words, such as "sex" (a strategy sometimes called "spamdexing").156 Different search engines use different strategies to deal with the problem,

none of them fully successful as yet.157 But a technological solution is probably on the horizon: it is the nature of this fast-developing technology that almost as fast as one form of cyber-mischief develops, electronic countermeasures will be found to solve the problem. It is equally probable, of course, that future cyber-pirates will find another way to make money off the Internet. By the time Congress passes a statute designed to deal with a problem, technology is likely to make the new remedy obsolete. Where the technology is not quite fast enough to solve the problem, courts, constantly confronting the latest version of cyber-piracy, can tweak existing doctrine and adapt. Congress has to go back to the drawing board each time, something that it is unlikely to do and that will ultimately be a waste of time.

And metatags are just the tip of the iceberg when it comes to technological innovations. Courts will soon face new copyright and trademark questions involving framing, hyperlinking, and keywords, all of which involve methods of getting from one web site to another without using a domain name.158 In the context of personal jurisdiction doctrines, addressed in Part III, courts will have to decide whether "cookies" and "packet-sniffers" are sufficient to confer personal jurisdiction.159 And if in the meantime Congress has passed a statute to govern personal jurisdiction on the Internet, I have no doubt that that statute will be at least as unhelpful in cookie and packet-sniffer cases as the ACPA has been in cybersquatting cases.

2. Remedial Innovations

In the case of cybersquatting, it is not only technology that has overtaken congressional enactments. Some changes have made cybersquatting less of a problem in particular cases. For example, the ongoing development of additional top-level domains (such as .biz for businesses, .aero for the airline industry, museum for museums, and name for individuals)160 will make conflict less likely: the child named Pokey will now be able to register pokey.name without interfering with the toy maker's desire to own pokey.com.161

However, the most significant development is the availability of an alternative to federal litigation by aggrieved trademark owners. In November 1999-ironically, the same month that the ACPA took effect-the Internet Corporation for Assigned Names and Numbers ("ICANN"), an international nonprofit organization with overall responsibility for domain names, established its Uniform Domain Name Dispute Resolution Policy ("UDRP"). Under the UDRP, any domain name registrant must agree to submit to nonbinding arbitration if its domain name is challenged by a trademark owner.

To prevail under the UDRP, a complaining mark owner must allege and prove that: "(i) [the registrant's] domain name is identical or confusingly similar to a trademark or service mark in which the complainant has rights; and (ii) [the respondent] has no rights or legitimate interests in respect of the domain name; and (iii) [the] domain name has been registered and is being used in bad faith."162 The UDRP provides a nonexhaustive list of three factors to determine whether the registrant has a legitimate interest in the domain name: (1) that the registrant used the name prior to notice of the dispute; (2) that the registrant is commonly known by the name; or (3) that the registrant is making a fair or noncommercial use of the name.163 The UDRP contains a further nonexhaustive list of four factors governing the determination of bad faith: obtaining the domain name (1) "for the purpose of selling, renting, or otherwise transferring it" for profit; (2) "in order to prevent the owner of the [mark] from reflecting the mark in a corresponding domain name"; (3) "for the purpose of disrupting the business of a competitor"; or

(4) to "attract, for commercial gain, Internet users ... by creating a likelihood of confusion . . . as to the source, sponsorship, affiliation, or endorsement of ... the web site."164

If these elements and factors sound familiar, they should. As one commentator has noted, the UDRP and the ACPA are "essentially the same law."165 But unlike the hastily enacted ACPA, the UDRP was several years in the making, and was adopted in response to both a U.S. Department of Commerce "White Paper" and a WIPO (World Intellectual Property Organization) Report, which in turn drew on successive drafts with extensive comment periods.166 And while the UDRP has been subject to some criticism, 167 it is apparently much more popular with trademark owners than is the ACPA: there are only about forty reported ACPA cases, but there have been over 4,000 arbitrations under the UDRP.168

The ACPA, then, offers a classic illustration of congressional regulation in an area that would have been better left to the judiciary and the international organization created to deal with exactly the problems that prompted the ACPA. Trademark owners prefer

the UDRP. The First Amendment and the judicial system are better served by continuing under the FTDA. The ACPA, on the other hand, benefits no one: it accomplishes little, and does so at some cost. Why, then, was it enacted? The next section returns to that question, raised first in the Introduction.

D. Congress and the Courts

Congress was not unaware of the ACPA's flaws; it was simply able to ignore them because they were so ephemeral. The legislative history of the ACPA offers a perfect illustration of what happens when one-sided pressure on Congress overwhelms its ability to take the long view or to consider intangible interests.

The ACPA was enacted as part of an omnibus appropriations bill. It was apparently considered so minor, in fact, that it was passed unanimously by both houses of Congress with little or no discussion.169 The President did not even mention the anticybersquatting portion of the appropriations bill in his signing statement.170

The only significant discussion of the ACPA occurred in a hearing before the Senate Judiciary Committee.171 That hearing exemplified the arguments against hasty passage of cyberlegislation. The committee heard from three live witnesses, each of whom also submitted a written statement.112 Those witnesses all represented trademark owners: one was the president of the International Trademark Association ("INTA"), another was a lawyer who represented trademark owners, and the third was the head of a company that investigated trademark violations on the Internet. All three stressed that cybersquatting was accelerating, and that it was confusing consumers and jeopardizing e-commerce.173 Their unani

mous opinion was that a federal anticybersquatting law was an immediate necessity.

But Congress was not left to ferret out potential problems on its own. Two law professors, both experts in intellectual property law, submitted written testimony. Their testimony presaged my arguments in this Article. Professors Jessica Litman and Michael Froomkin told Congress that the courts had been consistently ruling against cybersquatters under existing law, that WIPO was developing mandatory arbitration rules that would solve the problem, and that the proposed statute jeopardized First Amendment interests.174 Froomkin concluded that the bill "will do more harm than good."175

Congress was particularly made aware that the courts were taking care of the problem. Both professors provided detailed discussions of the existing case law: Litman stated that "courts have been merciless to defendants perceived as cybersquatters,"176 while Froomkin noted that "every cybersquatter who has gone to court has lost."177 The other witnesses could not refute this evidence. Instead, when asked about it, they replied that cybersquatting was on the rise and that cybersquatters were getting smarter and better able to evade existing trademark law.178 They provided no support for either assertion. Moreover, as my survey of the cases suggests, the courts uniformly dealt harshly with cybersquatters, however clever. Nevertheless, the Committee unanimously voted the bill out, and Committee members touted its virtues to the full Senate.

That Congress ignored law professors should come as no surprise,179 but there is a lesson here. Litman and Froomkin were

representing an abstraction, defending the judicial process and the First Amendment rather than anyone's concrete interests. The task fell to experts because these intangible public interests have no partisans. Congress did not enact this unnecessary law out of ignorance, but out of the political circumstances of the latest cybercrisis.

And the ACPA is not the only federal statute that creates more problems than it solves. The next section looks at E-SIGN, another unnecessary and potentially harmful federal statute governing cyber-technology. E-SIGN is neither as significant nor as problematic as the ACPA, but it provides another example of crisis-- induced legislation.

II. WRITING THE FUTURE: E-SIGN

If you are reading this Article online, are you reading a "writing"? Uncertainty about possible judicial answers to that question triggered concern about whether electronic contracts and electronic "signatures" would be held valid. The statute of frauds and other laws require certain contracts to be "in writing." Many contracts and other documents must be "signed" to be legally valid. The new technology used in e-commerce raised questions about the application of such requirements to contracts and signatures that were produced by-and often remained entirely within-the ones and zeroes of a computer's circuitry.

Courts have dealt with similar problems in the past, as other new technology supplemented or replaced paper-and-ink contracts. They have consistently held that telegrams and telexes constitute a writing sufficient to satisfy the statute of frauds180 and have assumed without deciding that faxes also do so.181 Any collection of symbols that the sender intends to serve as a signature does so, including names on telegrams (communicated orally to an opera

tor),182 telexes,183 faxes (whether originally handwritten or typed),184 and even simply a printed name and logo on a form.185

Beginning in the late 1980s, courts also held that information recorded on a computer disk constitutes a "writing" for various purposes.186 A few courts have now begun to hold that purely electronic messages even if not saved to a disk-are also "writings." In People v. Perry,187 for example, the court held that the electronic submission of false Medicaid claims was actionable under a statute penalizing the submission of false written instruments. And in Doherty v. Registry of Motor Vehicles, the court upheld as "signed" an e-mail message that identified itself as "the report of Trooper Thomas Kelley," and contained a statement that "[d]ata entry and transmission were done by Kelly, Thomas, by or at the direction of Trooper Thomas Kelley."188

Most recently, courts have upheld "clickwrap" agreements: before downloading software, the user must click a box that indicates his agreement to the terms and conditions specified by the software provider. Courts have uniformly held that clickwrap contracts satisfy all writing and signature requirements.189

Several commentators have thus concluded that neither a writing requirement nor a signature requirement is likely to stand in the way of electronic contracts.190 Businesses apparently agreed: the first paperless electronic real estate closing (banking and mortgages were previously one of the most paper-bound industries) took place before any federal legislation on electronic signatures was in effect.191

Despite this record of judicial accommodation of new technology, several state legislatures, the National Conference of Commissioners on Uniform State Laws ("NCCUSL"), and, eventually, Congress, decided that positive legislation was necessary to ensure that various writing requirements did not stand in the way of ecommerce. The result is the federal Electronic Signatures in Global and National Commerce Act ("E-SIGN").

According to one senator, E-SIGN "took three Senate committees 8 months and thousands of hours."192 Unfortunately, ESIGN has little to show for all that work. The statute has two primary provisions.193 First, no signature or contract may be denied legal effect solely because it is in electronic form.194 Second, any state statute regulating electronic contracts and signatures is preempted unless it mirrors the Uniform Electronic Transactions Act (UETA) recommended by NCCUSLI95 or unless it specifies alternative, E-SIGN consistent, technology-neutral procedures for the use

of electronic records or signatures.196 For its part, UETA is similar-although not identical-in primary substantive effect.197

Like the ACPA, E-SIGN was popular, especially with certain high-tech and financial industries. Also like the ACPA, however, ESIGN seems both unnecessary and unnecessarily complicated. It is unnecessary because, as described above, courts were well on their way to reaching the same substantive results under the common law. Moreover, by the time E-SIGN was enacted, some version of UETA had been enacted in eighteen states and was under consideration in eleven more.198 Several more states had enacted laws that regulated electronic signatures but that were not modeled on UETA.199 Thus, even more so than with cyber-piracy, Congress was acting in an area already adequately dealt with by other institutions.

It is certainly true, as advocates of E-SIGN argued, that not all state statutes-nor all judicial decisions--governing electronic signatures and contracts were identical. But e-commerce is no different from other interstate commerce insofar as it faces a welter of inconsistent and possibly conflicting regulations. Such is the price of our federal system; as Justice Brandeis recognized long ago, it has the compensating advantage of allowing state law experiments that can be copied by other states if successful.200 Indeed, in that sense the state statutes-whether or not modeled on UETA-offer an advantage over E-SIGN: like judicial decisions, individual state laws are more limited in geographic scope and more likely to be revised if necessary than are federal statutes.

Perhaps, given time, the courts and the state legislatures would have settled on a single successful formula for dealing with electronic signatures. If not, a federal statute enacted at that later date would have had several advantages over the hastily enacted ESIGN. Congress would have had before it a history of state legislative and judicial responses to the problems of electronic signatures; this history might have helped to eliminate unsuccessful ap

proaches and to identify potential problems. Waiting would also have allowed e-commerce to develop, both giving free rein to nonlegislative solutions and clearly identifying whatever problems remained.

The failure to await developments would not be a serious problem if E-SIGN itself did not create new legal problems. A statute that provides little benefit may be acceptable (although think of those three committees slaving away for eight months for so little gain!) if the cost is low. But E-SIGN, like the ACPA, seems to create more problems than it solves. Although there have been no cases to date, commentators have been quick to point out that the language of E-SIGN is far from clear.

The preemption provisions are particularly problematic. As noted earlier, E-SIGN allows state regulation of electronic records and signatures if a state adopts UETA or if a state adopts a law that is both technology-neutral and consistent with E-SIGN. As several commentators have pointed out, however, UETA is not entirely consistent with E-SIGN.201 Presumably, adoption of a "clean" version of UETA would nevertheless survive preemption analysis, although thereby reducing the national uniformity that was one goal of E-SIGN. But what if a state adopts a modified version of UETA? Would only the parts that are both different from UETA and inconsistent with E-SIGN be preempted, or would the adoption of such a modified UETA make all sections fair game for preemption on the grounds of inconsistency with E-SIGN?202

Other questions have also been raised. One of the largest differences between E-SIGN and UETA is in the consumer protection provisions, and there are conflicting suggestions in the legislative history of E-SIGN about whether a state adoption of UETA would displace those E-SIGN provisions.203 The requirement of technological neutrality leaves open questions about what to do when the authenticity of an electronic signature is called into question.204 The

federal Office of Thrift Supervision suggested that E-SIGN "has left some legal issues unresolved and, indeed, may have created new ones, particularly for online banking."205 Another commentator has said that "E-SIGN raises as many questions as it answers" in the health care context.206

All in all, E-SIGN, like the ACPA, may do more harm than good. As one commentator puts it, "At this point it is impossible to know whether E-Sign [sic] or the UETA will govern any discrete portion of any given transaction."207 That is a far cry from the certainty and uniformity that E-SIGN was supposed to deliver in order to promote e-commerce, but it is typical of a statute enacted without adequate attention to the existing landscape. E-SIGN is probably more benign than the ACPA, but it offers a further example of the phenomenon of Congress intervening in circumstances that should make us wary of the consequences.

If E-SIGN and the ACPA represent congressional failures, the arcane doctrine of personal jurisdiction illustrates the successes that can be accomplished when judges are left to develop the law through the common law method. The next section tells that story.

III. A COMMON LAw SUCCESS STORY: PERSONAL JURISDICTION AND THE INTERNET

It is hornbook law that a court cannot exercise jurisdiction over an unconsenting defendant unless the defendant has sufficient "minimum contacts" with the forum to make the exercise of jurisdiction consistent with "traditional notions of fair play and substantial justice."208 Unfortunately, the hornbook-and the clarity of the law-stops there. Courts continue to struggle, even in non-Internet cases, with questions about what activity constitutes minimum contacts and what circumstances make the exercise of jurisdiction fair and just.

The difficulty of determining whether minimum contacts exist is exacerbated when the primary contact with the forum is a web site that is accessible to anyone, including residents of the forum

state.209 Doctrines designed for cases involving physical objects sent into the state or even telephone calls into (or out of) the state are difficult to translate into an Internet context.210

Supreme Court precedent, muddy as it is, directs courts to examine whether the defendant has intentionally directed its activities toward the forum state or purposefully availed itself of the benefits of doing business in the forum state.211 In one sense, the operator of a web site is not directing its activities anywhere except toward the state in which the operator and the server are located-any transmission of information occurs only because a web surfer in another state "reaches out" for the information. Such a narrow reading of minimum contacts, however, would allow webbased companies to conduct business in every state but be amenable to jurisdiction only at home. On the other hand, one can argue that because a web site owner knows that its site is accessible to residents of the forum state (as well as of every other state), it has

intentionally directed its activities toward the forum state. But without more, this broad reading of the minimum contacts test means that any web site owner can be sued anywhere in the United States.212

A few early cases fueled a fear that application of the International Shoe test to the Internet would take the broad approach and eviscerate all due process limits on personal jurisdiction.213 Three 1996 cases from different district courts found personal jurisdiction on the ground that defendants had established web sites accessible to residents of the forum states.214 As a later court (rejecting the analysis) described it:

The courts in these three cases... chose to exercise personal jurisdiction for Similar reasons, which can be summarized as follows: through their web sites, defendants consciously decided to transmit advertising information to all Internet users, including those in the forum state, thereby (allegedly) committing trademark in

fringement in the forum state and purposefully availing themselves of the privilege of doing business with the forum state.215

Two state-initiated cases that same year also contributed to speculation that the Internet would massively expand courts' jurisdiction. In United States v. Thomas,216 the Sixth Circuit affirmed the conviction of two California defendants for maintaining a web site that violated the local community standards for obscenity in Memphis, Tennessee; the web site had been accessed by a government agent in Memphis. In State v. Granite Gate Resorts, Inc.,217 in which the Minnesota Attorney General sought to enjoin an out-of-state online gambling web site accessible to Minnesota citizens, a state court also found personal jurisdiction.

But within a year, courts had begun drawing distinctions that avoided both an overly broad and an overly narrow reading of International Shoe in the cyberspace context. The first-and still most influential-case to do so was Zippo Manufacturing Co. v. Zippo Dot Com, Inc.218 In Zippo, the court suggested that whether a web site accessible in the forum state conferred jurisdiction depended on "the nature and quality of commercial activity that an

entity conducts over the Internet."219 The court went on to identify three categories of web sites:

At one end of the spectrum are situations where a defendant clearly does business over the Internet. If the defendant enters into contracts with residents of a foreign jurisdiction that involve the knowing and repeated transmission of computer files over the Internet, personal jurisdiction is proper. At the opposite end are situations where a defendant has simply posted information on an Internet Web site which is accessible to users in foreign jurisdictions. A passive Web site that does little more than make information available to those who are interested in it is not grounds for the exercise of personal jurisdiction. The middle ground is occupied by interactive Web sites where a user can exchange information with the host computer. In these cases, the exercise of jurisdiction is determined by examining the level of interactivity and commercial nature of the exchange of information that occurs on the Web site.220

Since Zippo, courts have uniformly agreed that merely maintaining a passive web site is not sufficient to confer jurisdiction.221 To do otherwise, as many courts have recognized, would be to make the establishment of a web site tantamount to consenting to jurisdiction in every state.222 Moreover, as one commentator pointed out as early as 1998, finding jurisdiction on the basis of a purely pas

sive web site is inconsistent with current doctrines of personal jurisdiction, which require that the defendant "have consciously chosen to take advantage of a state's privileges and benefits."223

The Zippo analysis also avoids the opposite extreme, and thus prevents companies from doing a nationwide business on the web and nevertheless escaping jurisdiction in any state. Zippo itself involved a company that operated a web-based news service for subscribers. It had contracts with seven Internet access providers in the forum state and 3,000 individuals in the forum state. Each individual subscriber had submitted an application to the defendant containing his or her address, and, knowing that it was selling its service to a forum resident, the defendant had entered into the contract. Despite the fact that the defendant had no other contacts with the forum state, the court held that it had jurisdiction.224

Any other ruling would have taken an overly narrow approach to minimum contacts, allowing the differences between reality and virtual reality to provide a loophole from liability. One ingenious defendant, for example, suggested that although it entered into a contract with Virginia plaintiffs (after meeting them face to face in Virginia), and the plaintiffs performed the contracted work while sitting in front of computer screens in Virginia, there was no jurisdiction in Virginia because the work was actually "performed" in cyberspace. The court responded:

Although the defendants appear to be correct in their contention that much of the activity in this matter occurred in cyberspace, this can not signify that the increasingly large number of those who deal in e-commerce shall not be subject to jurisdiction in any earthly court. There being no District Court of Cyberspace, . . . [d]efendants will have to settle begrudgingly for the Western District of Virginia.225

Until there is a District Court of Cyberspace, then, courts faced with the two ends of the Zippo spectrum seem to have successfully adapted the jurisdiction doctrines of real space into the world of cyberspace.

The hard cases, of course, lie in the middle of the spectrum. Even here, however, courts have reached results that are, by and large, both arguably correct and increasingly predictable.226 Indeed, given the confusion that reigns in this area in non-Internet cases, the Internet cases are remarkably orderly. These intermediate cases can be divided into three categories and generally reach consistent results within each category.

1. Minimally interactive web sites. In some cases, the web site is interactive only to the extent that it allows the user to send e-mail-or other simple replies-back to the web site owner. Customers cannot place orders or otherwise do business over the web. Courts have uniformly found that such minimally interactive sites cannot, by themselves, provide sufficient contacts for the exercise of jurisdiction.227

Although language in some cases might suggest that minimally interactive sites can create jurisdiction, an examination of the cases themselves indicates that the web sites at issue were in fact more than minimally interactive. For example, in Thompson v. Hands-Lopez, Inc., the court focused on the fact that Texas residents could access the site, and it cited one of the three early personal jurisdiction cases that allowed jurisdiction on the basis of a

purely passive web site.228 But the web site involved in Handa-- Lopez was an online casino; customers had to purchase online gaming chips with a credit card before playing, and the web site "continuously interacted with the casino players, entering into contracts with them as they played the various games."229 I have found no cases-after the 1996 trio-in which a court exercised personal jurisdiction solely on the basis that the defendant maintained a minimally interactive web site.

Perhaps the most difficult case to date is GTE New Media Services Inc. v. BellSouth Corp.230 In that case, the plaintiff alleged that the defendants had conspired to monopolize the online business directory service by breaking the nation into regions to be served exclusively by each defendant and inducing Internet browser sites (such as Yahoo and Netscape) to link only to the authorized member of the conspiracy. Plaintiff alleged that the purpose of the conspiracy was to divert advertising business from plaintiffs web site to defendants'. Defendants' only contact with the forum (the District of Columbia) was that its web site could be accessed by forum residents. Although the site was interactive, the customer paid nothing to use defendants' search tools and entered into no transactions with defendants. The district court found jurisdiction on the basis of presumed effects in the forum, but the D.C. Circuit reversed.

As a formal matter, the site was highly interactive, transmitting information back and forth between a defendant's server and the consumer's computer. The court of appeals, however, appropriately refused to rely on such a wooden interpretation of the passive/interactive distinction. Since there was no commercial transaction occurring between the consumer and the defendants, the court held that the company was not transacting business in the forum.231 While this is a close case-the defendants were bene

fitting commercially from a web site accessible to, and evidently accessed by, forum residents-it is not clearly wrong (and may well be right). Moreover, if the case is ultimately found unpersuasive, it is unlikely to be followed by other circuits. GTE New Media Services thus illustrates one of the strengths of the common law method: the most important aspect of the case is the court's sensitivity to factual context in its refusal to apply the passive/interactive distinction in a formalistic manner.

2. Web sites that allow online ordering. In these cases, customers can place an order online (paying for it by providing a credit card number); the order is filled by the web site owner without any further communication. Many companies now have online ordering, including both e-businesses like amazon.com and traditional bricksand-mortar companies that also do a significant amount of off-line sales. The few cases involving true e-businesses have all found them to fall in Zippo's first category, easily holding that jurisdiction exists.232 But where a bricks-and-mortar company also allows online ordering, the issue is more difficult. Online ordering capability is similar to a passive web site insofar as the site owner has no control over who has access to the site, but it differs from a passive web site to the extent that a transaction can be consummated online.

Courts have recognized both the similarities and the differences between passive web sites and sites allowing online ordering. Whether a court will find jurisdiction over a web site owner on the basis of an interactive site that allows online ordering depends on two factors: (1) is specific or general jurisdiction sought; and (