This work is licensed under a Creative Commons Attribution 4.0 International License.
The following article was first published, in a shortened version, in the February 2000 issue of the journal Copyright World. It has also appeared in Computer Law Reporter, Vol. 31, Number 1, March 2000, and Internet Law and Business, Vol. 1, Number 5, March 2000.
The 1990s have witnessed the creation of an entire new category of intellectual property—the collection—as well as a new (sui generis) right of ownership. In this article I will try to summarize the issues that content providers and their representatives should be alert to when dealing with laws concerning “collections of information”, a term I will use interchangeably with the term databases.
To understand the importance of this issue, you must sense how deeply collections of information have penetrated the life of economically advanced cultures. This morning, you may have flipped through a listing of television channels to find one where you could check the weather for your region; you may have grabbed a bus schedule on the way to your office and (if you are calorie-conscious like many Americans) peeked into a book that rates food’s nutritional value before choosing a breakfast item from a restaurant menu; you may then have settled into your computer and called up a database of statistics critical to your profession. Not only the statistical database, but also the TV guide, the weather listing, the bus schedule, the nutrition book, and the restaurant menu are all collections of information.
The EC Databases Directive1 defines a protected collection as “a collection of independent works, data or other materials arranged in a systematic or methodical way and individually accessible by electronic or other means” (Article 1, paragraph 2). One current U.S. Congressional bill (H.R. 354) calls it “information that has been collected and has been organized for the purpose of bringing discrete items of information together in one place or through one source so that persons may access them.” Both definitions are very broad, being drawn up to exclude only individual works covered by copyright.2 And with so many databases offered over the Internet for easy searching and retrieval, we have become a society awash in collections of information.
My own perspective on intellectual property encompasses the roles of both consumer and producer. I work as an editor at the major computer book publisher, O’Reilly & Associates (but this article expresses solely my own opinions). I am not a lawyer, although during the course of writing this article I have consulted more lawyers than you would wish to gather around a single table. For the past five years I have also intensively researched many policy issues related to information creation and use for a public-interest group I volunteer for, Computer Professionals for Social Responsibility.
I feel that the first concern of content providers is the scope and purpose of protection laws; therefore I have ignored such details as enforcement and transition periods. Furthermore, I have concentrated on U.S. legislation because that is where I have the richest sources, but I have been able to indicate broadly the scope of European law.
Database manufacturers base their call for a new right on purely economic grounds, unlike existing forms of intellectual property that are grounded philosophically on the promotion of creativity, or “moral rights” in the European tradition. An author tends to maintain the copyright on a creative work even when he or she receives no remuneration for it (as in the case of this article) simply so other people won’t change it and ship it around in garbled form. But database manufacturers have little reason to be concerned about how people use facts from the collections unless the manufacturers’ markets are threatened.
Due to the economic basis for protection, the EC Directive and related laws call for penalties only where the original manufacturer has made a “substantial investment” in the database. Furthermore, the manufacturer must show that copying has harmed its market (to which the proposed U.S. legislation adds “potential markets”). On the other hand, the economic basis suggests that, to uphold the public interest, regulators and courts should balance database protection with an insistence on fair licensing and other promotion for competition.
There is practically no research that would bolster the economic argument for special database protection law. Since the database industry is growing rapidly and enjoying strong profits, no one convincingly demonstrates that it is harmed by the lack of special laws. Instead, justification for a sui generis right is usually limited to statements that seem obvious and require no defense.
Collections made through a substantial investment deserve protection from copying.
Technological change facilitates free-loading (the resale of information that another entity has paid to collect).
New rights for collection makers will improve their markets and rates of return.
Even in Europe, where proponents have succeeded in passing a database right into law, it is worth re-examining these foundations in order to understand where courts should limit the right.
The doctrines of “unfair competition” and “misappropriation” have long been used to protect database manufacturers, like other providers of goods and services, from losing business to competitors who become “free riders” by simply republishing the work that the original manufacturer has taken so long to acquire.3 But the doctrines have always been tempered by a recognition that the material being “appropriated” in this case is data, whose ready availability is critical to informed public decisions in science, politics, and numerous other fields. Where there is a monopoly on the collection of data, the effect on public access can be onerous; Reichman and Samuelson cite, for instance, the case of Landsat satellite images, which, when privatized, went up in price from $400 to $4,400 per image.4
Modern opportunities for value-added databases—also known as “transformative uses5”—heighten the risks of protecting existing investment. Astounding amounts of new value can be created by “reformatting, filtering, and hot-linking”6 data from various sites. Sometimes, perhaps, restrictions on copying can encourage different database creators to share data that would otherwise remain secret—for instance, when realtors combine their listings7—not only because they are guaranteed license fees, but because they can prevent misuse of private information. But such a listing doesn’t require database protection; trade secrets and contracts should be adequate. Courts should be alert in balancing the manufacturer’s right to a market or “potential market” with the value of new products from second comers.
Because a huge amount of data can be copied from an electronic medium or an Internet site in the blink of an eye, database manufacturers argue that copying will become ubiquitous in the absence of laws. True, devices that make CD-ROMs can now be bought for a few hundred dollars. But copying a CD-ROM is copyright infringement, because the expression as well as the data is being reproduced.
Database protection may be appealing because it can provide an extra bulwark against copying. The technological argument is not as unassailable, however, as it appears. Some technological advances favor the manufacturer, not the copier: a combination of encryption, digital watermarks and electronic signatures can theoretically restrict the number of copies made or mark copies as illegal. These technical measures—also called “self-help measures” because they allow the manufacturer to prevent infringement without legal intervention—have admittedly not been perfected. But simple anti-copying mechanisms already appear in video (DVD format) and music (SMDI format). We may be entering an age when, as law professor Lawrence Lessig claims, “copyright is more effectively protected than at any time since Gutenberg…In such an age—in a time when the [technical] protections are being perfected—the real question for law is not, how can law aid in that protection? but rather, is the protection too great?”8
Fears that technical progress will cause illegal copying to proliferate are not unique to our time. Ginsburg notes such concerns going back to the improved printing presses of the 1880s. Some 20th-century courts extended protection for compilations, but others moved in the opposite direction, stressing the need for copyrighted works to display an author’s “personality”.9
Both Peters10 and a statement by Tyson and Sherry11 warn that inadequate legal protection may discourage manufacturers from offering their wares in “convenient” formats such as CD-ROM. Both were writing before the advent of high-bandwidth data telecommunications, which are gradually decreasing in price throughout the technologically advanced world and which make online access much more attractive than before. The possibility that the Internet (where data can be updated instantly) will nearly eliminate disks as the medium of choice shows the risks of making laws in a technologically fast-moving field. On the Internet, a server can detect enormous amounts of copying from a single user and block access to that user.
While some opponents of database protection laws look to technical self-help measures as alternatives,12 intense debates in other areas of intellectual property have revealed that these measures have socially negative implications and open the door to egregious abuses by content providers:
Encryption and license enforcement mechanisms can eliminate traditional fair-use types of protection.13 If they limit the number of elements that can be extracted, they hinder research. If they prohibit the copying of media, they eliminate archiving, with severe impacts on libraries. And unlike legal protections, they include no expiration date. (Any policy, including an expiration date, could theoretically be built into the technology, but each such complication increases cost and the risk of failure.)
Digital watermarks and digital signatures can harm privacy by tracking and preserving information on users’ habits.
Thus, we should heed database protection proponent Ginsburg when she says, “Privatizing information through contract, encryption, and similar devices may carry greater individual and social costs than would a copyright system.”14 The British database protection law, the CDPA, offers a nod toward the risks of self-help measures by overruling licenses that would remove (the few remaining) rights that the law reserves to users.15
In an ideal world, the balance between measures protecting manufacturers and measures promoting competition would give manufacturers the resources they need to develop new products, while leaving alive the incentive to develop them before competitors grab market share. In the real world, greed erects legal measures that reward sloth. It is all too easy for protected manufacturers to keep prices high and use a legal club to hold back competition. The results would be not only less research and less innovation, but also less utilization of the service because it is overpriced and underfeatured. Manufacturers may end up hogging pieces of a small pie when they could have enjoyed a much bigger one.
Since all sides agree that data is the lifeblood of modern society, legislators and courts must carefully balance the incentives to produce databases (such as protection) with the freedom to use data in new ways and create value-added or transformative products. Policy decisions must rest partly on how often competitors can reasonably be expected to regenerate the original data on their own.
Proponents and critics of database protection disagree about how widespread monopoly (“sole-source”) databases are. Reichman and Samuelson cite high start-up costs, difficulty differentiating a new entrant from a well-known established provider, the inability of niche markets to support multiple vendors, and other barriers to entry.16 I would add to this list the situations where an organization generates data for internal use, funding it to benefit other projects, and then offers the data to the public as an ancillary source of revenue. For instance, the American Medical Association (which testified in support of data protection) maintains a Physician’s Masterfile containing “demographic, educational, certification, licensure, and current practice information for over 800,000 active U.S. doctors of medicine (MDs) and over 90% of the doctors of osteopathic medicine (Dos), including members and nonmembers alike”17 which they obviously can use for numerous organizational purposes. While the facts are by no means secret, no other organization could conceivably reach and query all the same physicians in order to build up a competing database.
Arguing the other side of the question, Tyson and Sherry claim that, where no artificial monopoly exists, competition will naturally arise to keep any one company from charging too much or refusing outright to license data.18 This argument seems to assume perfect markets in a classical economic sense.
A frequent proposal by those trying to temper protection laws—and even some in favor of such laws, such as Ginsburg in a relatively early 1992 publication19—is to require compulsory licensing for databases, carried out through arbitration or rates pre-set by law.20 While this may risk bringing bureaucracy into a fast-moving technological field, the principle has popped up repeatedly in many contexts. Thus, the first proposal for an EC Directive (April 15, 1992) required “fair licensing” for sole-source databases. The final Directive, however, steered away from such direct intervention and simply said a “right must not be used in such a way as to facilitate abuses of a dominant position.”21 Note that “a dominant position” is not necessarily limited to a sole-source provider. Since licensing is not regulated, the lesson I draw from the discussions is that courts must apply anti-monopoly and pro-competition laws vigilantly to prevent abuse of database law.
All authors admit that, to some extent, data produced under government auspices should be exempt from protection, because a sole-source situation is created by definition. Many commentators in the U.S. (including several members of the Clinton Administration) insisted that the exclusion extend to anything produced by federal government funding or even “within the scope of” a federal agency; consequently, this strong exemption appears in the current versions of both U.S. bills.
Finally, the locking down of data in protection laws can be tempered by an expiration time, as in copyright law. The EC Directive calls for protection to expire after 15 years. Many commentators have pointed out that expiration times are mostly irrelevant for online databases because these are usually updated continuously, with old and new data intermingled in such a way that users cannot determine what data is new and what is old; the upshot is permanent renewal of the ban on copying.22
The landmark U.S. Supreme Court ruling, Feist Publishing, Inc. v. Rural Telephone Service Co.,23 established that mere “sweat of the brow” did not endow collections of information with copyright protection. Instead, “creativity in selection or arrangement” was required. Later cases have gone to great length to interpret those words. The case of Feist was fairly stark because Rural Telephone’s “selection” included all customers and their “arrangement” was a standard alphabetical listing. The defendant, Feist, needed to copy the data in order to put out a unified telephone listing covering several companies; Rural Telephone refused to license the data while the other companies had cooperated.
Friends of database manufacturers disagree strongly with their critics over whether Feist represented a departure from previous precedent24 or a clarification of precedents that had been in place for decades.25 Even Ginsburg, who has argued repeatedly that the Supreme Court “dropped a bomb” with Feist,26 and who has examined some 200 years of U.S. legal history to show protection for collections of information, admits that early American case law reserved only “right of reproduction, but not the right to make derivative works” to publishers of compilations.27
The wording in the ruling indicates that the Supreme Court saw no earth-shattering precedent in Feist, but only a clarification of what the U.S. Constitution has said all along. However, it is really irrelevant whether Feist was new, because it was perceived as new throughout the database community. Even an article pointedly titled “What’s All the Fuss About Feist?” concludes by saying, “Feist has clearly altered the landscape of copyright protection of factual compilations.”28 Not only has Feist been cited by every article and testimony I’ve seen in relation to database protection in the U.S., but it became a shot heard ’round the world, turning up in the first call for database protection laws (save for some “experiments” in Nordic countries29), the April 15, 1992 proposal for an EC council directive.30
Countries vary in how much originality is required for traditional copyright protection on collections. Rees, for instance, says that “the standard of originality in UK law is considerably lower than the standard required in many other EU countries, notably Germany.”31 In his opinion, copyright is not limited to expression, but to an “elaboration or aggregation of information” once it is set out in an “original work”.32 A later passage in the book adds that “since early nineteenth century, UK law” has protected works that are “original” in the sense that “sufficient skill, labor, or experience must have been applied in the production of the work.”33 The word “original” here is clearly different from its use in the U.S. Feist ruling because no “creative spark” is required.
Despite the breadth of copyright protections in some countries, the EC could not have added extra protection to databases just by expanding copyright law because Article 2(5) of the century-and-a-half-old Berne Convention clearly defines essentially the same limits as the U.S. Supreme Court on copyright for collections: “collections which, by reason of the selection and arrangement of their contents, constitute intellectual creations.”34 Similarly, the World Intellectual Property Organization Copyright Treaty, Article 5, extends protection to “compilations of data or other material, in any form, which by reason of the selection or arrangement of their contents constitute intellectual creations.”
Because of Feist’s historical centrality, I cannot bring myself to leave it behind without pointing out what a poor poster child Rural Telephone makes for the proponents of database protection. Were the Feist case brought up again today under current database laws, Rural Telephone would lose again.
First, Rural Telephone fails the “substantial investment” clause; it does not require much sweat of the brow to record the telephone number assigned to each customer who signs up for service.
Second, Rural Telephone possessed its listing because of government protection as a monopoly telephone company. The EC Directive and database laws exclude government data, although the Directive is much less clear than the proposed U.S. legislation in defining government data. Whether or not a court might interpret the “exclusively licensed by” government exclusion to cover the activities of a telephone company, it was the government that gave Rural Telephone an exclusive right to gather the information Feist needed.
Third, Rural Telephone was preventing both competition and innovation, deliberately refusing to license the data in order to hold back a value-added product. As several commentators have pointed out, Feist was no free rider. And all sides on the database protection question agree that it must be balanced by anti-competitive regulation, supported by court action.
The model for anti-competitive actions in data is the cases brought against the publisher Magill in Ireland by three major television broadcasters.35 Magill TV Guide Ltd published a comprehensive directory of television programs available to Irish viewers, taking its data from the three broadcasters’ listings. Sued by all three, it was forced to stop publication by an Irish court. Magill took the cases to the EC, which ruled that Article 86 of the EC Treaty provided an exemption from the usual copyright laws because the broadcasters were abusing “a dominant position.” They also praised Magill for its added value in publishing “a new product in the form of a multi-channel guide”. While settled in the public interest, these cases raise the chilling question: would Magill be able to win today in the face of database protection laws? And even if it could, how would a small researcher or innovative start-up firm marshal the resources to take a case all the way to the EC?
Later sections of article will show that database protection laws provide rich environments for abuse of monopoly positions; courts, regulators, and defendants must all be vigorous in applying anti-competitive measures in return. The EC Treaty allows courts to award exemptions from laws in order to promote competition.36 But it is not clear to me that the exemptions can be invoked to protect value-added products; perhaps the “Refusal to supply” abuse can be invoked.37
The few cases after Feist in the U.S. that offer precedents for database protection tend to show that new laws are unnecessary. In Key Publications, Inc. v. Chinatown Today Publishing Enterprises, Inc.,38 Kregos v. Associated Press,39 and Bellsouth Advertising & Publishing Corporation v. Donnelley Information Publishing, Inc.40 the collections were all held to be under copyright protection. (National Basketball Association v. Motorola, Inc.41 is irrelevant to the discussion because the defendant discovered its data independently from the plaintiff.) One could argue in response, perhaps, that Feist discouraged many manufacturers from suing.
The most disturbing post-Feist case is Warren Publishing, Inc. v. Microdos Data Corporation,42 which dangled back and forth in the courts, but where the court finally ruled (with three dissenters) that cable listings were not original enough to deserve protection. One lawyer critical of database protection laws argues that the court simply made a mistake.43 A representative of technology companies further points out that Warren could have taken action under the doctrine of misappropriation, but brought defeat on itself by its failure to do so.44 My own reading of Warren suggests a very confused court majority that thought that Warren’s choices displayed creativity in the “discovery of facts” rather than creativity in arrangement. As nerve-racking as it may be for a database manufacturer to watch millions of dollars of investment turn on a phrase, this case is not enough to show the need for a whole new property right. And now that the principle of originality is clear, it should be trivial for database manufacturers to make sure that some “spark” of originality is included in their products.
Reichman and Samuelson, tracing the history of the EC drafts, point out that the EC did not originally envision a sui generis right as demanded by database manufacturers. They apparently intended a modest extension of the “unfair competition” concept to cover more collections.45 Even the concept of misappropriation can be, so to speak, misappropriated to create an anti-competitive situation in data, as pointed out by Gordon.46 But misappropriation is a time-honored concept with ample precedent, and has a built-in defense when the defendant can claim that the database manufacturer is a sole source and has refused to deal fairly. Even Ginsburg, who is extremely solicitous of database manufacturer’s needs, reacted to Feist by calling merely for Congress to pass a law concerning misappropriation.47 However, proponents of a sui generis right in the U.S. have complained that other solutions—notably misappropriation and contract law—vary too much from state to state.48
Reichman and Samuelson tell a sad story of a reasonable EC initiative that descends over time to become a final Directive that gives manufacturers a far-reaching and highly dangerous sui generis right with weak balancing clauses. Reichman and Samuelson’s account backs up the frequent claim, with which even Rees “would not argue”, that the Directive was “a cynical attempt to boost the European information technology market at the expense of the rest of the world.”49 They further claim that it has effectively banished the concept of “incremental” innovation that “is central to the development of modern technological paradigms” and that the EC Directive “harbors no working conception of a public domain whatsoever.”50
The first proposal for an EC Directive protected all non-commercial users by saying that “unfair extraction” was defined only in the context of commercial use.51 Stripping out such protection, however, the final Directive instead inserted an exemption in Article 6, paragraph 2(b) for “teaching and scientific research”. The clause was weakened by being made optional, and if national laws include it, it must apply only to non-commercial use.
The EC Directive does not protect explicitly against independent discovery of the facts in an existing database. The U.S. bills, by contrast, include a paragraph protecting the “gathering or use of information obtained by other means”.
Another landmine in the EC Directive is the assignment of rights to “lawful users” of a database instead of simply “users”. While “lawful use” is fairly straightforward to define in physical media (because the medium is either sold or explicitly licensed) it can be abused by Internet sites. The landmine may explode when Web sites try to impose restrictions on the use of their data. For instance, several sites (Shetland Times Ltd. in Scotland, Ticketmaster in the U.S., and several U.S. newspaper publishers suing Total News52) have tried to force users to view material by starting at the top-level “home page” and have sued sites that make “deep links” to material below. Sites wishing to control access will now try to claim, perhaps, that only users coming down through the home page are “lawful users.”
How much copying is to be considered infringement? Proponents of legislation insist there is no hard and fast lower limit.53 The EC Directive criminalizes any reproduction, “in whole or in part.”54 One U.S. bill, H.R. 354, softens this almost Levitican insistence on purity by using the word “substantial” to modify copying: database protection extends to “substantial” portions of the database as well as to copying the whole. But U.S. commentators state that “substantial” must be considered qualitatively as well as quantitatively, and there is no clear minimum amount of copying that can be defined as “substantial”.55 The result of such reasoning is that users extracting information have no guidelines at all as to how much is too much, as interpreted by a court.
The EC Directive on the Legal Protection of Databases was passed on March 11, 1996 with an implementation deadline of January 1, 1998. According to Band,56 12 members of the EC have passed laws to comply with the Directive, and the other 3 are being sued to force compliance. The Directive includes a reciprocity clause that tries to force other countries to pass similar laws by threatening to favor EC companies over outsiders whose countries lack database protection.
Flush with success at the EC, database vendors persuaded the U.S. House of Representatives to consider a database protection law and got the U.S. delegation to the World Intellectual Property Organization to propose an international treaty. Database protection has foundered on both fronts, however. WIPO refused to take action on a U.S. proposal after informational meetings in 1996 and has never taken up the issue since. That the pre-eminent organization defining intellectual property was satisfied to leave database protection at the existing level, refusing to pass EC-style protection, should leave grave suspicions in the minds of content providers concerning database protection laws. Critics ascribe the failure in both WIPO and the U.S. to the increased consciousness and organization of the research, computing, library, and public-interest communities, along with technology providers.
Furthermore, other countries outside the EC are slow to respond to the reciprocity clause. Some Latin American countries have passed laws.57 But Reichman58 states that Japan has chosen to extend unfair competition law instead. According to Prue Adler59 of the Association of Research Libraries, many underdeveloped countries opposed the draft WIPO database treaty because the industrialized world’s existing investment in collections would put it in a significantly advantageous position.
Critics of database protection in the U.S. see the reciprocity clause as something of a paper tiger, because collections can be organized to meet copyright criteria and be protected under standard copyright law.
The U.S. Congress has considered no less than four versions of a “collections of information antipiracy act” during the past four years; H.R. 354 is the current version. To the credit of its sponsors, they have progressively tempered it to meet criticism, often incorporating the suggested wording of its critics verbatim. They have moved in the reverse direction of the EC, starting with a sui generis right and moving closer to a combination of unfair competition and misappropriation.60 Recently an alternative to H.R. 354 has been introduced: the “Consumer and Investor Access to Information Act”, H.R. 1858.
If a repackaging of the data can be considered a new market, its copying is safe from the type of database protection set up by the EC Directive. But the U.S. bill H.R. 354 adds “potential markets” (a phrase with precedent in copyright law61) as areas where a database manufacturer can prohibit copying. Thus, a database manufacturer can prohibit a competitor from copying data by showing the court that it is preparing to provide products to the competitor’s market—not a comparable product, perhaps, but just another product which the competitor might cause customers not to buy.
The “potential market” phrase, which was originally left undefined, gave manufacturers the right to stop innovation simply by saying, “Oh, yeah, we could do that if we wanted.” Current wording requires the manufacturer to “reasonably expect to derive revenue, directly or indirectly” from the potential market, or to have taken “taken demonstrable steps to offer in commerce within a short period of time a product or service”. The loopholes left by these definitions do not offer much assurance to competitors. H.R. 1858 is far preferable because it considers a copy to be infringing only if it is a “duplicate” of the original, defined as “substantially the same”. Thus, value-added and transformational uses are protected. Furthermore, H.R. 1858 defines two criteria that must both be met before the bill applies: that the copy “displaces substantial sales or licenses” of the original and that the copy “significantly threatens the opportunity to recover a return on the investment” of the original.
The provisions comparable to fair use in copyright law have expanded over the years, as commentators argued that the EC Directive’s exemption for non-commercial teaching and scientific research was wholly inadequate. While even the Registrar of Copyrights dismissed the exemptions in the first U.S. bill as serving “a primarily symbolic value”,62 H.R. 354 exemptions now cover extraction for:
illustration, explanation, example, comment, criticism, teaching, research, or analysis—but only “if it is reasonable under the circumstances.” Several clauses follow that try to define what is “reasonable,” but it would take too long to dissect them here.
nonprofit educational, scientific, or research purposes—but only if it “does not materially harm the primary market for the product or service”, a clause that effectively removes the protection just granted.
news reporting—this time with no qualifications to weaken the exemption.
genealogical information for nonprofit purposes.
government collections—a broad clause including contractors and any one funded by the government or working under the “scope” of the government (with small exceptions).
computer programs—a complex clause that may leave loopholes, but is supposed to limit protection to the contents of the database rather than the process of accessing it.
digital online communications—a very specific exception for databases used in routing and other networking functions.
Proponents of H.R. 354 see this long list as an assurance that all socially valuable activities have been protected. Critics would not only point out loopholes, but would ask how many more socially valuable activities may be discovered to be under threat from database law after it has been in effect for a few decades. Or as Gordon writes: “the community may never recognize the loss of new works chilled by the misappropriation explosion.”63
H.R. 1858 provides much more safety for copying for research purposes, protecting it “so long as such conduct is not part of a consistent pattern engaged in for the purpose of direct commercial competition”. It also includes exemptions for government information and computer programs.
I have shown that the case history suggesting the need for database protections laws is just as “thin” as the constitutional protection cited by the Supreme Court for databases. But I have found two hypothetical examples cited by supporters of protection for collections. In particularly, they are brought out to support the contention that, while everyone agrees the amount of copying must be “substantial” before it can be called infringing, there is no way to specify a minimum amount that triggers the “substantial” clause and consequent legal penalties of H.R. 354. But the hypothetical examples raise other important questions about database protection as well.
The Physician’s Desk Reference is a weighty hardcover tome that lists manufacturers’ information on medications available in the United States along with their uses, counter-indications, side effects, etc. It is found in every clinic and many other human service centers, but there is no conceivable reason a nonprofessional would buy one unless he suffered from a serious case of hypochondria. A U.S. House report on the 1998 database protection bill cites the following hypothetical example as behavior they plan to criminalize:
If a second comer extracted information about the thousand most commonly prescribed medications and offered it for sale to the general public—for example under the title “Drugs Every Consumer Should Know”—that extraction and use, although a fraction of the total collection of information, would cause the kind of market harm that Committee intends H.R. 2652 to prevent.64
The House committee cites this hypothetical example to show importance of prohibiting partial extractions of databases. But the example shows even more the importance of maintaining competition, and of keeping up the pressure on the original manufacturer to expand offerings, in order to improve the quality of life and the public availability of critical information. If “Drugs Every Consumer Should Know” would be of interest to ordinary consumers, and if the publisher of the PDR has not written it already, I think that reuse of the data by a competitor (whether copied from the PDR or assembled from the original manufacturers) would be a marvelous idea.
In testimony before Congress on a alternative database protection bill, Andrew J. Pincus of the U.S. Commerce Department cited the following hypothetical case to show why the bill must leave a loose definition of how much copying is too much:
For example, the book edition of a national database of hotels might subdivide hotels by state and city; in such a situation, we understand that the intent [of a provision in the database bill] would be to create liability when a competitor misappropriated all of the Sacramento, California or Cincinnati, Ohio listings from the national database, even though this might only be a small part of the national database.65
The collections of restaurant listings I’ve seen usually show originality in arrangement, which ought to qualify them for copyright protection. If the restaurants were listed by city, county, or postal code, to be sure, there would be no originality. But many listings use more vaguely defined geographical breakdowns that are useful to both residents and tourists, such as “Historic Section” or “Riverside District”. Unless such areas are formally defined, they represent choices that can be used in court to justify copyright protection.
This example again shows that computers and the Internet make possible an infinite variety of organizations for the same data to serve different users in different ways. If extracting a geographically limited set of restaurant data and including it in a larger listing (say, along with stores and museum) hurts the market for the original collector of information, that is completely the fault of the original collector for failing to exploit an opportunity. The definitions of H.R. 1858 protect competition. The “potential market” definition in H.R. 354, while suppressing the competitor who copies data, is supposed to assure that the public will get the service soon from the original collector, but it is uncomfortably broad.
I have given the database protection proponents their shot at justifying new laws; I now have the right to hypothesize in order to show the risks of such laws.
A first hypothetical example is based on a real-life case involving Web sites that offer online auctions. Two such sites, AuctionWatch.com and Bidder’s Edge, allow people to find the best possible deals by searching for items through multiple auction sites—in other words, AuctionWatch.com and Bidder’s Edge translate a user’s search into a search on each competing site, combine the results of those sites, and present the user with Web links to all matching items. The added value (combining information from multiple competitors) precisely parallels both the Feist case and the Magill case, where courts upheld the right to copy material in order to produce a combined listing.
The best-known Web auction site, eBay, complained repeatedly about AuctionWatch.com’s and Bidder’s Edge’s display of its listings, and finally in November 1999 sued Bidder’s Edge. Because the courts have not resolved the copyright status of links to Web sites, it is hard to predict the outcome. People technically well versed in Web technology tend to see links as references, and therefore subject to the fair use exemption. But in the disturbing Shetland Times case, the court ruled that unauthorized linking was infringing. Related U.S. cases were settled out of court on vague grounds where the linking site paid a license fee to the site where the page resided.
A future Ebay/ Bidder’s Edge case might exploit database law to claim that a combined listing from multiple sites is an infringing copy of data. The state of disorganization in laws regarding the Web suggests an additional danger. A user must download data from a Web site to his or her own computer in order to view it, and this temporary downloading has been considered a “copy” by several courts. (And there is some justification for the rulings, because anything loaded into temporary computer memory can easily be stored permanently.) A future eBay-type plaintiff might claim that anyone approaching its site through a combined listing on another site is “unlawful” under the scope of the EC directive. Not only that, but the site creating the link has engaged in contributory infringement. (I hope that I have not suggested a course of action to any unscrupulous lawyers.) If viewed favorably by the courts, such definitions could bring the innovative services for which the Web is so famous to a screeching halt.
Another hypothetical example returns to the incredible power of computers and the Internet to create new value by “reformatting, filtering, and hot-linking” existing data.
Suppose that a chemical firm releases databases listing the concentrations of various chemicals in the soil and water of many communities. A medical firm lists various diseases broken down by community. A third firm extracts data from the chemical and medical sites to produce detailed reports correlating chemical contamination with concentrations of disease.
While the public benefit of such a project goes without saying, both H.R. 354 and a database protection law conforming to the EC directive would allow either the chemical site or the medical site to quash the reports. Under H.R. 354, the site would have to make the claim that they were planning to offer similar reports soon, or that people buying the third firm’s report would otherwise buy the original databases. But the chilling effect of the bill might prevent anyone from trying to offer a new service in the first place.
Early critics of database law pointed to the needs of researchers to use large amounts of data. Tyson and Sherry disputed the risk, claiming that scientists and researchers always “compress” their data during legitimate research so that the original facts cannot be reproduced.66 But in this example, preservation of the original data is critical—nobody could formulate environmental or medical policy from a report that combined data to say, for example, that a broad class of chemicals were correlated with disease in general. The example underscores Reichman and Samuelson’s warning, “Any legal or policy initiative that could weaken the ability of second comers to enter and compete effectively in markets for products that add value to existing data should, therefore, be carefully scrutinized, lest they impede competition without offering benefits to the public.”67
A sui generis right to collections of information is now embodied in law throughout the European Community. The United States has held off from passing such a law for three years, and may still choose among three alternatives: maintaining the status quo, passing H.R. 1858 with its strictly limited injunction against “duplicates”, or passing H.R. 354, which is philosophically similar to the EC Directive but includes more exemptions to protect competition and the public interest.
Every content provider is also a content consumer. My own field, technical book publication, is like maple syrup production near my home in New England: you need 50 liters of sap from the tree to make a single liter of syrup. Even a highly creative offering like a motion picture or a novel requires an enormous amount of research. Thus, companies in a content production industry should worry more about the free flow of source material than in building fences around their products. They should study database protection laws and the defenses against them, particularly defenses based on competition. In this way they hopefully can balance protection for the liter of maple syrup they produced this Spring with their ability to tap another 50 liters of maple tree sap next Winter.
I would like to thank Prue Adler, Jonathan Band, Nicholas Bohm, and Susan Singleton for answering my questions, and Prue Adler, Jonathan Band, and Karen Coyle for reviewing a draft of this paper. I take responsibility for any errors. Finally, I must thank the New England School of Law, a small private institution, for allowing the public into their library, where I performed the research without which I could not have written the paper. Their staff took a sincere interest in my project and spent a good deal of time digging up source material for me. Libraries stand as the great inspirations for those of us who believe in the open dissemination of information to the public.
Andrew Oram works as an editor at O’Reilly & Associates, a major publisher of computer books, software, and services of interest to computer users. He also used to follow issues of Internet policy for Computer Professionals for Social Responsibility. His policy-related publications are available at https://www.praxagora.com/andyo/professional/article.html and this paper is at http://www.praxagora.com/andyo/professional/collection_law.html. He once participated in the making of maple syrup when he accompanied his children on an educational excursion. The opinions in this paper are his alone.
1 http://www2.echo.lu/legal/en/ipr/database/text.html
2 For instance, see the statement of Marybeth Peters, Registrar of Copyrights, before the House Subcommittee on Courts and Intellectual Property on H.R. 354, March 18, 1999. http://www.house.gov/judiciary/106-pete.htm
3 Gordon, Wendy J., On Owning Information: Intellectual Property and the Restitutionary Impulse, 78 Virginia Law Review 149-281, 1992.
4 J.H. Reichman and Pamela Samuelson, Intellectual Property Rights in Data?, Vanderbilt Law Review, Vol. 50, No. 1, January 1997, pp. 51-166., p.121.
5 See, for instance, the statement of Joshua Lederberg, President-emeritus, Rockefeller University, on behalf of the National Academy of Sciences, National Academy of Engineering, Institute of Medicine and American Association for the Advancement of Science before the Committee on the Judiciary, U.S. House of Representatives, 18 March 1999. http://www.house.gov/judiciary/106-lede.htm
6 J.H. Reichman and Pamela Samuelson, op. cit., p. 125.
7 Testimony of the National Association of Realtors before the House Judiciary Subcommittee on courts and Intellectual Property on H.R. 354, the “Collections of Information Antipiracy Act,” presented by Terry McDermott, Chief Executive Officer and Executive Vice President, National Association of Realtors, March 18, 1999. http://www.house.gov/judiciary/mcdermott.pdf
8 Lessig, Lawrence, Code and Other Laws of Cyberspace, Basic Books, New York, NY. 1999.
9 Ginsburg, Jane C., Creation and Commercial Value: Copyright Protection of Works of Information, Columbia Law Review, Vol. 90, No. 6, pp. 1865-1938, 1990, p. 1893.
10 See, for instance, the statement of Marybeth Peters, Registrar of Copyrights, before the House Subcommittee on Courts and Intellectual Property on H.R. 2652, October 23, 1997, http://www.house.gov/judiciary/41112.htm
11 Tyson, Laura D’Andrew and Sherry, Edward F., Statement for the Information Industry Association on Economic & Policy Issues in Statutory Protection for Databases before the Committee on the Judiciary, October 23, 1997. http://www.house.gov/judiciary/41118.htm
12 For instance, see the statement concerning H.R. 2652 before the Subcommittee on Courts and Intellectual Property, Committee on the Judiciary, House of Representatives, Washington, D.C., October 23, 1997 by J.H. Reichman, Visiting Professor of Law, University of Michigan and Professor of Law, Vanderbilt University. http://www.house.gov/judiciary/41121.htm
13 See, for instance, Tyson and Sherry, op. cit.
14 Ginsburg, op. cit., p. 1921.
15 Brazell, Lorna, Limitations and Exceptions, in Rees, Christopher, and Chalton, Simon, eds., Database Law, Jordans Publishing Ltd, 1998, pp. 74-75.
16 Reichman and Samuelson, op. cit., p. 94.
17 Statement of the American Medical Association to the Subcommittee on Courts and Intellectual Property of the Committee on the Judiciary, U.S. House of Representatives re: H.R. 2652 "Collections of Information Antipiracy Act", presented by Richard F. Corlin, M.D., February 12, 1998. http://www.hyperlaw.com/ama.htm
19 Ginsburg, Jane C., No “Sweat”? Copyright and Other Protection of Works of Information After Feist v. Rural Telephone, Columbia Law Review, Vol. 92, pp. 338-388, 1992, p. 1918.
20 For example, see Reichman and Samuelson, op. cit., or Lederberg, op. cit.
21 Chalton, Simon, EU Directive 96/9/EC on the Legal Protection of Databases, in Rees, op. cit., p. 19.
22 For instance, see Chalton, Simon, EU Directive 96/9/EC on the Legal Protection of Databases, in Rees, op. cit., p. 23, Rees, Christopher, Database Right, p.66 in the same work, Reichman and Samuelson, op. cit., p. 90.
23 Feist Publishing, Inc. v. Rural Telephone Service Co., 499 US 340 (1991).
24 See, for instance, Tyson and Sherry, op. cit.
25 See, for instance, response by Pamela Samuelson re: Tyson/Sherry Report, October 23, 1997. http://www.arl.org/info/frn/copy/psamlet.html
26 For instance, see Ginsburg’s statement before the Subcommittee on Courts, Intellectual Property and the Administration of Justice, October 28, 1997, H.R. 2652, http://www.hyperlaw.com/ginsburg.htm
27 Ginsburg, Jane C., Creation and Commercial Value: Copyright Protection of Works of Information, op. cit., p. 1878.
28 Sheils, Paul T., and Porchina, Robert, What’s All the Fuss About Feist? The Sky is Not Falling on the Intellectual Property Rights of Online Database Proprietors, 17 Univ. Dayton Law Review, 563-585 (Winter 1992), p. 585.
29 Reichman and Samuelson, op. cit., p. 74.
30 Chalton, Simon, EU Directive 96/9/EC on the Legal Protection of Databases, in Rees, op. cit., p. 17.
31 Rees, Christopher, Information and the Law, in Rees, op. cit., p. 6.
33 Chalton, Simon, Copyright in Databases and Other Compilations, in Rees, op. cit., p. 73.
34 Chalton, Simon, EU Directive 96/9/EC on the Legal Protection of Databases, in Rees, op. cit.
35 Radio Telefis Eireann v. Commission of the European Communities, Case T-69/89 [1991] ECR, The British Broadcasting Corporation and BBC Enterprises Ltd v. Commission of the European Communities, Case T-69/89 [1991] ECR, and Independent Television Publications Limited v. Commission of the European Communities, Case T-76/89 [1991] ECR, Case T-69/89 [1991] ECR, all published in Reports 1991-7/II.
36 Trevor Cook, Licensing and Enforcement, in Rees, op. cit., p. 90-91.
37 Trevor Cook, op. cit., p. 120.
38 Key Publications, Inc. v. Chinatown Today Publishing Enterprises, Inc., 945 F. 2d 509 (2d Cir. 1991)
39 Kregos v. Associated Press, 937 F. 2d 700 (2d. Cir. 1991)
40 Bellsouth Advertising & Publishing Corporation v. Donnelley Information Publishing, Inc., 999 F. 2d 1436 (11th Cir. 1993)
41 National Basketball Association v. Motorola, Inc., 105 F. 3d 841 (2d. Cir. 1997)
42 Warren Publishing, Inc. v. Microdos Data Corporation, 115 F. 3d 1509 (11th Cir. 1997)
43 Testimony of Jonathan Band on behalf of the Online Banking Association before the Subcommittee on Courts and Intellectual Property of the United States House of Representatives Committee on the Judiciary on the Collections of Information Antipiracy Act, H.R. 2652, February 12, 1998. http://www.hyperlaw.com/band.htm
44 Testimony of Tim D. Casey on behalf of the Information Technology Association of America before the U.S. House of Representatives Subcommittee on Intellectual Property and the Courts concerning H.R. 2652, 12 February 1998. http://www.hyperlaw.com/casey.htm
45 Reichman and Samuelson, op. cit., p. 84.
47 Ginsburg, No "Sweat"?, op. cit., p. 367.
48 See, for instance, Summary of Statement of Ms. Marilyn Winokur on Behalf of the Coalition Against Database Piracy on H.R. 354, the Collections of Information Antipiracy Act, before The House Judiciary Subcommittee on Courts and Intellectual Property, March 18, 1999. http://www.house.gov/judiciary/106-wino.htm. See also the testimony of Robert E. Aber, Senior Vice President and General Counsel, The Nasdaq Stock Market, Inc., on behalf of the Information Industry Association before the Subcommittee on Courts and Intellectual Property, House Committee on the Judiciary Hearing on H.R. 2652, the Collections of Information Antipiracy Act, February 12, 1998. http://www.hyperlaw.com/aber.htm
49 Rees, Christopher, Database Right, in Rees, op. cit., p. 64.
50 Reichman and Samuelson, op. cit., pp. 89-90.
51 Chalton, Simon, EU Directive 96/9/EC on the Legal Protection of Databases, in Rees, op. cit., p. 21.
52 Shetland Times Ltd. v. Wills (1997) S.L.T. 669. Ticketmaster Corp. v. Microsoft Corp. (California District Court, unreported, 1997). Washington Post Company v. Total News Inc. 97 Civ 1190 (PKL) S.D.N.Y. For these citations and an excellent analysis of these cases I am indebted to Morrison, Alex, Hijack on the road to Xanadu: The Infringement of Copyright in HTML Documents via Networked Computers and the Legitimacy of Browsing Hypermedia Documents, Journal of Law and Technology, 1999, Issue 1, February 26, 1999. http://www.law.warwick.ac.uk/jilt/99-1/morrison.html
53 For instance, see Marybeth Peters, Registrar of Copyrights, before the House Subcommittee on Courts and Intellectual Property on H.R. 2652, October 23, 1997, op. cit.
54 Chalton, Simon, Copyright in Databases and Other Compilations, in Rees, op. cit., pp. 55-56.
55 Statement of Marybeth Peters, Registrar of Copyrights, before the House Subcommittee on Courts and Intellectual Property on H.R. 2652, October 23, 1997, op. cit.
57 Testimony of Daniel C. Duncan, Vice President, Government Affairs, Software & Information Industry Association, before the Subcommittee on Courts and Intellectual Property, House Judiciary Committee Legislative Hearing on H.R. 354, the Collections of Information Antipiracy Act, March 18, 1999. http://www.house.gov/judiciary/106-dunc.htm
58 Statement concerning H.R. 2652 before the Subcommittee on Courts and Intellectual Property, Committee on the Judiciary, House of Representatives, Washington, D.C., October 23, 1997 by J.H. Reichman, op. cit.
60 Statement of Marybeth Peters, Registrar of Copyrights, before the House Subcommittee on Courts and Intellectual Property on H.R. 354, March 18, 1999, op. cit. See also Lederberg, op. cit.
61 Statement of Marybeth Peters, Registrar of Copyrights, before the House Subcommittee on Courts and Intellectual Property on H.R. 2652, October 23, 1997, op. cit.
62 Statement of Marybeth Peters, Registrar of Copyrights, before the House Subcommittee on Courts and Intellectual Property on H.R. 354, March 18, 1999, op. cit.
64 Quoted in a memorandum from an attorney for the U.S. Department of Justice on Constitutional Concerns Raised by the Collections of Information Antipiracy Act, H.R. 2652. http://www.acm.org/usacm/copyright/doj-hr2652-memo.html
65 Prepared Statement of Mr. Andrew J. Pincus, General Counsel, Department of Commerce, before the Subcommittee on Telecommunications Trade & Consumer Protection on regarding H.R. 1858, June 15, 1999. http://com-notes.house.gov/cchear/hearings106.nsf/768df0faa6d9ddab852564f1004886c0/2951e93fc0523370852567900055716c?OpenDocument
67 Reichman and Samuelson, op. cit., p. 125.
This work is licensed under a Creative Commons Attribution 4.0 International License.