Legal doctrine often evolves both more slowly than the conduct and the technology it is intended to regulate and by analogy. When parties argue that previously unheard-of behavior should generate some legal response, the relevant institution (generally Congress, the courts, or administrative agencies, such as the U.S. Patent and Trademark Office) often seeks guidance from past experience. Earlier decisions addressing a specific factual context give some confidence that applying the legal rules from these prior situations to the new fact pattern will produce a sensible result.
But principles developed to deal with one set of facts can only imperfectly fit another. This imperfect fit often becomes more evident over time, sometimes raising the question of whether the initial analogy itself was valid. Legal doctrine may then evolve away from its first response to a new condition toward more specific rules reflecting an increased appreciation of the points at which the original analogy does not hold.
For example, the U.S. courts that first addressed claims of software copyright infringement in the 1970s and 1980s and Congress, when it decided to include computer programs within the category of "literary works" under the Copyright Act, analogized writing computer programs to writing novels. However, over time, the courts began to have difficulty using this analogy to apply rules developed with books and manuscripts in mind to software. In particular, long-established law held that a novel’s plot, if sufficiently detailed, is protected by copyright. Courts found it difficult to translate this "old" law into the new context of computer programs in which the "plot" is much more difficult to see and understand. Additionally, market realities demand that other programs connect to the initial copyrighted one as, for example, Lotus 1-2-3 and Excel each run on top of Microsoft’s Windows. Moreover, any PC operating system that hopes for any chance of marketplace success must run at least the set of applications run by Windows, the dominant system. Achieving compatibility may necessarily result in similarities between the "plots" of the original and the compatible programs. Should a court hold the author of the later-developed program an infringer?
Initially, the courts granted fairly broad protection to program structure.1 Over time though, as they began to appreciate how computer programs differ from traditional literary works and to understand the economics of the software industry, they developed new rules specific to software to define the extent of copyright protection for a program’s "plot."2
While software copyright cases continue to be difficult to adjudicate, the law has at least evolved to some measure of stability. The Internet is now the legal frontier on which the process of setting rules by analogy and adjustment is just beginning. Courts have begun to grapple with questions like who may access a Web site, in what manner, and for what purpose. Their resolution of these issues will greatly influence the structure of competition on the Internet for the foreseeable future.
They represent a difficult challenge, because the law has never before dealt with rights of access to a medium like the Internet or "property" like a Web site. But courts have historically addressed such questions as who may enter another person’s land or use another’s chattel (personal property that is tangible, but, unlike land, also portable) under trespass law. This time-tested legal doctrine, at least at first glance, seems to provide an apt analogy for adjudicating legal claims based on unauthorized access to and use of Web site data.
Closer analysis reveals, however, that an unthinking transplantation of tangible property law to the Internet is likely to create some undesirable results. This article explains why Web sites seek to limit access to and use of their information, why the principles of real and tangible personal property law present attractive analogies, and why use of these prior rules of decision is inappropriate without adjustment. It concludes by proposing some factors courts should consider in evaluating claims whose gravamen reduces to charges of unauthorized access to and use of Web site data.
Limiting Access to Web Sites
The notion of limiting access to Web sites may seem odd to those who recall the Internet as a research network where open access and the free exchange of scientific information was the norm. It was, however, inevitable, that as commercial firms began to turn to the Internet for revenue, they would begin to view their Web sites as valuable "real estate" to be protected against those who would use their existence or content to make money for themselves. Indeed, while the reasons a site owner may wish to limit access to its site are as numerous as Web sites themselves, a large percentage may be classified under the broad heading of preventing "free-riding" by others that threatens the visited site’s revenue model in one way or another.
The most obvious example of a site that would literally have to limit access to continue in business is one that relies on subscription revenue for its existence. If nonsubscribers are legally permitted to circumvent the subscription and payment process, the site’s primary revenue stream is dissipated. Similarly, some sites provide certain data for free but impose a charge for access to additional information. If visitors can avoid the fee and obtain the "restricted" information anyway, the site’s revenue stream disappears.
Other, perhaps less obvious examples, exist. Regardless of their revenue model, some sites may limit access to certain of their parts to users who have met certain qualifications. For example, a porn site may limit access to interior pages to those providing some assurance that they are 18 years of age or older.
Many sites continue to generate at least part of their revenue from advertising, often charging others for the privilege of placing banner ads on their home pages. The advertising rates are usually set with reference in some way to the number of "hits," or visits to the home page.3 If another site links directly to interior pages, the visited site may lose revenue, because no hits to the home page are recorded. Sites may be reluctant to take technological steps to block deep linking because the linker may circumvent the block. The linker’s practical ability to overcome virtually any technological barrier may lead to a "technological arms race" in which the site continually builds higher fences only to have the linker knock them down. Indeed, the increasing number of lawsuits involving complaints about deep linking evidences sites’ unwillingness to rely solely on technological solutions.
A site may also object to deep linking because it takes content out of context. The site may have been designed with a particular layout and intended viewing path. By coming in through a "side door" rather than the front, the deep linker sees the content out of order and in a way that might be misleading.
The deep-linking example illustrates how difficult it is to divorce questions of limiting access from defining the permissible means of access. Most sites would not argue that they can prevent another from employing a hyperlink to their home pages. It is the manner of accessing the site through a deep link to which they object.
In some of the most recent cases, the manner of access under attack is a bit less visible to the naked eye. Internet users have employed search engines throughout the medium’s brief history as a commercial marketplace. Many search engines contain relatively limited content of their own. Instead, they use software programs called spiders or robots to crawl the Web, copying information about others’ Web sites in the course of building their own databases. When a user enters a query, the engine searches its database and returns, minimally, the URLs (addresses) of sites likely to contain content of interest to the user. Historically, the indexed sites have had no reason to object to a spider’s activities. Their revenue models almost always depended in some way on the number of visitors to their sites. Thus, the more travel to a site, the better—whatever the source, whether from a search engine, a link on another site, or the user’s entering the URL into a Web browser directly.
An unthinking transplantation of tangible property law to the Internet is likely to create some undesirable results.
But this "win-win" situation in which the indexed sites welcome search engines and users employ them as a primary means of locating information has begun to break down. Brands and brand names are establishing themselves on the Internet. The more powerful the brand, the more likely users will know its URL and travel directly to the site. An established site may prefer that users not access a search engine from which they may obtain information about competitive offerings and choose to visit another site.
Additionally, spiders, like all visitors to a site, impose a burden on the sites they visit, using system resources that would otherwise be free to process other requests, especially those from potential customers. At least theoretically, a large number of spiders crawling a site could cause the system to slow down or even crash. Either event can erode the site’s goodwill with consumers and other paying customers. In the worst case, the performance degradation could drive a site out of business if it cannot either stop the spidering legally or technologically or afford to purchase (or legally compel the spider to purchase) additional capacity. In reality, however, the worst case scenario seems unlikely to occur. An individual spider likely imposes a minimal burden,4 and Internet economics will not support an unlimited number of spiders. The advertising dollars many search engines depend on for their existence are limited, making it unlikely that a large number of spiders would both search and cripple a site.
In evaluating claims of unauthorized access, it is also difficult to ignore the eventual use of the information in which an unauthorized visitor might choose to engage. The visitor might wish to copy the visited site’s information for personal or directly competitive purposes or as raw material in offering a new product. For example, one might copy eBay’s product and pricing information as part of a research project on auction sites. One might copy it in the course of developing a directly competing auction site. A site like Bidder’s Edge might copy it as part of the process of offering a new product, such as an index into auction sites that aggregates the search process. A user can go to Bidder’s Edge, enter a search query, such as "Derek Jeter autographed baseball," and receive a list of all of the auction sites offering the item and their prices. Bidder’s Edge is simply a specialized search engine. Other uses are likely to be in a gray area, potentially competing with the original site but also offering a new product that adds value to the original.
Intuitively, some of these uses offend less than others but all free-ride on eBay’s content in some way. The question for policymakers is how to distinguish permissible from impermissible free-riding as a legal matter. The intertwined nature of the questions of access, legitimate means of access, and allowable uses of the information so obtained makes the legal response especially difficult to formulate.
Why Trespass Is an Attractive Analogy
Since time immemorial, common law courts have sustained a cause of action in trespass to real property. Trespass to real property occurs whenever a person enters another’s property without the owner’s consent and without a legally recognized privilege that allows entry in the absence of such consent. It makes no difference whether or not the intrusion causes harm to the property or to its owner. Moreover, because the law recognizes few privileges, a landowner’s rights in the U.S. in his or her property are close to inviolable.
Various rationales support this broad right to exclude. Landowners often have some financial and psychic attachment to their property, especially if they use it for personal rather than business purposes. A landowner may view any physical intrusion, however harmless, as threatening, and be tempted to respond with force and expend resources to keep out unwanted visitors. That intruders will be liable for trespass both deters their conduct and makes it less necessary for the landowner to resort to violence or to erect a barrier to entry.
Additionally, landowners are more likely to invest in improvements if they have a right to exclude others from entering their property and taking advantage of these improvements for free. A strong, certain right to exclude allows landowners to safeguard the value of their property by giving them the right to refuse entry, without conditioning that right on the construction of a fence that may be costly to build. This rule is further justified by other cost considerations. There are many would-be trespassers but usually only one or a few parties holding rights in the land, and the boundaries of the land itself are readily ascertainable. It is cheaper for those who wish to enter the land to approach the "well-known" landowner for permission than to require landowners to seek out and negotiate with those whom they cannot easily identify but want to exclude from entry.
Trespass to chattels applies the principles of trespass to land to personal property. It adjusts the real property cause of action to account for the differences between land and movable property. A person may be liable for trespass to chattels whenever he or she intentionally takes possession of another’s chattel or causes a physical contact with or otherwise uses it. Unlike trespass to land, trespass to chattels requires the plaintiff to prove harm before being allowed to recover damages through a legal proceeding. The law justifies these differing rules by reference to the relative strength of the plaintiff’s interest in the inviolability of his or her property. The law regards the attachment to land and interest in safeguarding it against intrusion as stronger than in the case of chattels.
Web sites are personal rather than real property. They are not land, and the servers on which they exist are generally freely portable. Also, the information a Web site contains is easily transferrable through the transmission of electronic signals. Therefore, the law on trespass to chattels that regulates the use of another’s personal property seems an appropriate analogy in assessing claims of unauthorized access to Web sites. Visiting and using a Web site is the virtual analogy to touching and using a chattel.
In the 1990s, U.S. courts first applied trespass to chattels in the electronic context in cases involving spam (unsolicited bulk email).5 Analogizing trespass to real property cases, they held the spam’s electronic signals to be sufficiently tangible to interfere with the system receiving them. The legally cognizable harm arose not from damage to the chattel itself or a decline in its value, as in traditional cases, but from the impairment of the system’s value. If it is occupied in processing spam, it cannot be processing other, desired communications. System providers may also suffer damage in the form of canceled subscriptions and expenditures to install technical measures to filter out spam. Additionally, consumers may be harmed because they incur costs in disposing of the unwanted email.
The analogy to conduct like deep linking and unauthorized spidering is straightforward. The entry through a "side door" by the deep linker and a spider’s crawling are both unwanted virtual contacts with a Web site. The deep link causes damage by decreasing the linked site’s revenue. Like spam, spidering decreases the system’s availability for desired visitors. This performance slowdown may cost the site goodwill and, ultimately, revenue, because fewer visitors get to the site. To counteract performance degradation, the site may have no choice but to spend money to purchase more capacity or upgrade its filtering systems. An application of a trespass to chattels theory to address deep linking and spidering could be based on an analogy to the spam cases. Of course, spam cases are themselves based on an analogy to traditional, physical-world trespass to real property cases.
Why Trespass to Chattels May Not Be an Appropriate Cause of Action
Trespass to chattels seems, at least on the surface, to provide a workable analogy for the law to use in addressing claims of unauthorized access to Web sites. The differences between the physical and the virtual worlds, though, suggest that the law should develop a cause of action that takes into account the particular policy interests and characteristics of the Internet.6
Unlike land and chattels, a Web site is intangible property containing only information. The public interest in access to information on Web sites is likely to be greater than its interest in accessing another’s land or chattels. There are First Amendment free-speech concerns that should be weighed before the law grants rights to control access to Web sites.7 Trespass does not account for such interests.
Historically, copyright law has provided the set of rules governing exclusive rights in information, balancing these rights with the First Amendment rights of users. Notably, it has never accorded the copyright owner the exclusive right to control the manner in which a user views the copyrighted work. There is no law that prevents a person from starting a book in the middle, even though it might be misleading to take the material out of context in this way. There is also nothing to stop people from fast-forwarding past the commercials for which sponsors have paid when they watch a tape-recorded TV program. That the deep linker enters in the middle thus does not seem particularly troubling from a copyright perspective.
The question for policymakers is how to distinguish permissible from impermissible free-riding as a legal matter.
The copyright law has had less to say about access, leaving that question to other law. For example, a person might make "fair use" of a passage in a book. That this use would be permitted by copyright law does not entitle the person to break into a locked room to obtain a copy of the book. State law generally forbids this impermissible access. That the law respects a padlock—even one that protects access to information—in the real world suggests it should also do so in the virtual one. But publicly available Web sites are not locked in any sense, virtual or otherwise.
Indeed, Congress, in enacting the Digital Millennium Copyright Act in 1998, addressed the question of which access controls should be respected. Under the Act, "[n]o person shall circumvent a technological measure that effectively controls access to a [copyrighted] work."8 A technological measure effectively controls access when it "in the ordinary course of its operation, requires the application of information, or a process or a treatment, with the authority of the copyright owner, to gain access to the work."9 The provision is intended to protect against hacking or cracking systems that require some form of digital key for entry. By inference, Congress is less concerned about circumvention of measures that do not provide digital information with as high a level of protection against intrusion. It would indeed be anomalous if state law on trespass to chattels were interpreted to prevent access to Web sites that federal law would allow. It would be particularly disturbing if trespass to chattels were to prevent access to uncopyrighted information when federal law (with which state law may not conflict) would permit access to copyrighted data.
The policy basis supporting trespass does not translate well into the virtual context. The physical separation between the actual entities conducting deep linking or spidering and the owner of a visited Web site renders the possibility of physical violence quite unlikely. The interest of Web site owners in the inviolability of their sites also seems significantly less than that of tangible property owners. The Internet has a tradition of open access to Web sites that have not deliberately erected a barrier to entry. A firm opts into this system merely through the act of placing its site on the Web.
A Modest Proposal
While trespass to chattels seems to have little to do with the policy interests or realities of the Internet, there are probably cases in which the law should allow a Web site owner to limit access. For example, the site’s owner should usually be able to keep out those who have not paid the customary price. The owner should also be able legally to enjoin access conducted by means that place an unreasonable, unexpected burden on system resources. What is an "unreasonable" burden may vary depending on the facts but has yet to be implicated in actual litigation. A site should be able to stop copying and using its data when the copier is simply a pirate adding no new value. The key is to balance the incentive to create the initial Web site with incentives to develop new technologies like spiders and new products like the auction aggregator service offered by Bidder’s Edge. Trespass does not allow this balancing.
A better test would be one that allows a court to consider a number of factors, including: the means and purpose of the access; the nature of the information taken; the burden the unwanted access places on the visited site; the effect of the access on the visited site’s revenue; the cost to the visited site to gather the information it offers; and whether the access threatens the visited site’s existence or quality. Such a test should also consider the public benefit that might reasonably be expected from allowing access to continue, as well as whether the visited site has taken technological measures to block access. The law should be more willing to enjoin unwanted access when the site has invested in technological measures like those protected under the Digital Millennium Copyright Act. Under such a test, a site like Bidder’s Edge that offers consumers comparative pricing information without unduly burdening the eBay site would likely be allowed to continue its conduct. In contrast, a different site that copied information intending to offer it in direct competition with the visited site would often be enjoined.
A balancing test like this creates some uncertainty, especially when compared to a bright-line rule that would allow all site owners to exclude whomever they want for whatever reason. Additionally, unlike a strong property right to exclude, such a test would tolerate a technological arms race until the site invests in measures protected under the Digital Millennium Copyright Act, and possibly beyond. However, at this relatively early stage of the Internet’s development as a commercial marketplace, it may be appropriate to err on the side of weaker property rights that allow new, value-added technologies to flourish. The law may move to a regime of stronger rights over time, but, at least at the outset, weaker rights may enable a more vibrant network to develop.
Conclusion
The questions of who may access a Web site, how they may do so, and what they may do with the information they so obtain are not likely to go away. Courts today are facing pressure to apply law developed in the real world to the Internet by analogy. This approach risks obscuring the salient differences between the real and the virtual worlds. At this early stage of the technology, the more appropriate approach would be to follow the example of the software courts but to do so quickly—by recognizing these differences as the first claims arise and creating a new test appropriate to the Internet’s special characteristics.
Join the Discussion (0)
Become a Member or Sign In to Post a Comment