RELEASABLE X

NOT RELEASABLE

EXAM NO. ________

Fall 2003

This examination consists of a single essay question. You have 24 hours after picking up the exam from the Registrar to compose your answer (preferably typed or printed from a computer) and return both the exam and your answer to the Registrar.

You are permitted to use or refer to any materials you like in composing your answer. However, you may not cite as authority for any proposition of law anything that is not found in the textbook, any class handout materials, your class notes, or the materials found at the Course Materials page of the class web site. In other words, if you know of a case that we did not study but that you think is dispositive of an issue, you will get no points for saying "Under the principle of A v. B, issue 1 must be decided in favor of X." However, you may use the reasoning of A v. B and apply that, preferably with your own improvements to the fact situation in the problem. In my opinion, an A++ answer can be written based solely on the materials that we have studied, and I would advise wasting no time looking for other cases or materials.

GOOD LUCK!

Problem

Environmental Data, Inc. (EDI) markets to state governments a computer program/database system called Ecofacts. Ecofacts consists of a computer program called Ecoprogram and a related database called Ecodatabase. The system allows state government officials to visit manufacturing plants anywhere in the state and record data, using Ecoprogram to store the data in Ecodatabase. Ecodatabase is organized into various fields (categories) that have been carefully chosen to meet the needs of state environmental regulators in, among other things, assessing fines for overpolluting, taxes and tax credits based on use of certain kinds of equipment and the degree to which various pollutants are emitted, and estimating air and water quality. EDI does not collect any data itself. Rather, it licenses Ecoprogram to state governments. The state government officials themselves enter the data into their computers. The entered data is sent over the internet to an EDI central computer housing Ecodatabase. Assume (unrealistically) that EDI’s licenses with state government agencies do not expressly forbid making the information in Ecodatabase available to others nor do they expressly forbid reverse engineering Ecoprogram. (This assumption is necessary to avoid contract and copyright misuse issues that we did not study fully enough to warrant examination upon.) EDI claims a copyright in the Ecofacts system and asserts that such copyright covers both Ecoprogram and Ecodatabase. EDI also claims rights to Ecodatabase under H.R. 3261, which you may assume has become law in the form we studied in class, codified in title 17 as a new chapter of the Copyright Act. (Do not get into the question of whether this statute is constitutional.)

The Ecofacts system (Ecoprogram and Ecodatabase together) is designed to allow state regulators to manipulate the data contained in Ecodatabase to meet their needs, including printing out isolated single data entries and the results of calculations based on data taken from Ecodatabase. However, the program has no feature allowing the user to print out Ecodatabase in its entirety or to make an electronic copy of Ecodatabase. While an expert programmer who examines the Ecoprogram source code could readily devise a way to obtain a full printout or electronic copy of Ecodatabase, the average user would find it very tedious and time-consuming to do either operation completely and accurately.

Life Quality, Inc. (LQI) is a nonprofit entity whose mission is protecting the environment from pollution. It collects environmental data wherever it can and has its experts analyze the data for the purpose of writing reports and engaging in legislative lobbying and other political activity aimed at stronger environmental-protection regulation. LQI sought to license Ecofacts from EDI, but the price was more than LQI’s limited budget would allow. LQI then requested a state environmental protection agency to give it access to Ecodatabase for the purpose of copying its factual content, which LQI would completely reorganize to meet its own needs and purposes. The state agency refused, asserting fears that to do so might involve copyright infringement. (Ignore the 11th Amendment difficulties associated with this assertion.) LQI then brought an action under the state open records law, which allows public access to data and information generated by state funds upon paying the cost of making the copies (but provides an exception for copyright-protected works). LQI offered to pay the cost of reverse engineering Ecoprogram, if necessary, to get an electronic copy of the data. That suit is still pending. Meanwhile, EDI has brought an action against LQI seeking to enjoin EDI’s [sic, should read "LQI’s"] open-records suit and against LQI’s proposed extraction activities, arguing that, if LQI does acquire the data contained in Ecodatabase via the state agency, (1) it will infringe EDI’s copyright, (2) it will infringe EDI’s rights under H.R. 3261, (3) it will violate the DMCA, (4) it will constitute a trespass to chattels, and (5) it will violate the Computer Fraud and Abuse Act.

Evaluate EDI’s claims.

We know that, under Feist, EDI has no copyright rights in the factual data. Judge Posner confirms this on similar facts in his recent Assessment Technologies decision. The only potential copyright EDI can have is in EDI’s selection of the categories into which the state officials place the environmental data. Judge Posner seems to assume that selecting these categories evinces sufficient originality to qualify for copyright protection. But, as we discussed at length in class, protection of category selection in a nonelectronic database (e.g., telephone yellow pages) is a questionable basis for copyright protection, because of user friendliness, standardization, and efficiency considerations (functionality). Baker v. Selden argues that intellectual property protection should be sought under patent, and not copyright, law.

The argument is even stronger in the case of an electronic database, because protection of the categories (fields) amounts to protection of a data structure. A data structure is a nonliteral element of a computer program. Indeed, it is part of the "certain result" that execution of the "statements or instructions" constituting the program brings about. A data structure does not exist as an independent work. Rather, it comes into existence only upon an implementation of the structure’s concept into program code. The code, in other words, produces the data structure. The code is protected, but if different code brings about the same structure (result), it should be deemed noninfringing. This same answer can be achieved doctrinally by observing that data structures are chosen for efficiency (and perhaps compatibility), and not aesthetic, reasons. As nonliteral elements of a computer program, they should be filtered out by the Altai filters. Moreover, even as independent works, data structures are no different from the menu command hierarchy found unprotected as a "method of operation" in Lotus v. Borland. We have also seen that data structures are protectable under patent law if they meet the more stringent requirements of that regime. For this reason, too, there is no reason to stretch copyright to cover them. Still, many courts have been, or appear to be, willing to protect data structures under copyright, so we must continue the analysis on the assumption that EDI does have some sort of copyright in its categories.

From the facts, it appears that LQI will not be taking the data in the form in which it is organized in Ecodatabase, or if it does take it in that form (perhaps because as an initial matter that is the only way it can be extracted efficiently), it intends wholly to reorganize it. EDI played no role in the selection of the individual pieces of data that are stored in Ecodatabase; these were all determined and supplied by state regulatory personnel. If LQI does completely reorganize the data, its end product will not make use of the protected element of the database owned by EDI, namely, the selection of the categories in which the information is held. This means that the only possible infringement can come from copying of the database in the form designed by EDI as an intermediate step in producing LQI’s noninfringing ultimate product. Here, we may safely follow Judge Posner’s application of Sega, which held that making even an exact copy of a protected work for the sole purpose of extracting unprotected information to be used to create a noninfringing work was a fair use. LQI will not be marketing Ecodatabase to EDI’s customers, nor will it use Ecodatabase itself for any of the purposes for which EDI markets the database to its customers (fair use factor 4). Its purpose (fair use factor 1) is to create a new work that disseminates the unprotected information in a new form, something that copyright law should encourage, not discourage. The nature of the work (factor 2) also favors fair use: Protection for compilations is "thin," and that should hold true even where protection is limited to the selection of categories, because of their functionality. While here, we may assume, LQI did copy the entire protected element in its intermediate stage (factor 3), that appears to be necessary in order to get at the unprotected data, to which EDI can make no claim to a property interest. Copying of the entire work was also a factor in Sega, but the court properly found a fair use. A similar fair use analysis will apply to the extent LQI must reverse engineer Ecoprogram in order to allow it to extract the factual content of Ecodatabase.

In short, there is no copyright infringement because (1) EDI has no copyright rights in Ecodatabase or (2) if it does have copyright rights, the protected element (creative original selection of categories) was not copied, or (3) if the protected element was copied, such copying is a fair use.

The statute defines a database as a "collection of a large number of discrete items of information produced for the purpose of bringing such discrete items of information together in one place or through one source so that persons may access them." The data in Ecodatabase would seem to satisfy this criterion. There are many pieces of data, they are "discrete" in the sense that each data point is a different fact, and they are in one source for the purpose of easy access and manipulation. We may assume that LQI’s activity is "in commerce" (because its reports surely cross state lines) but does or will LQI "make available" a "quantitatively substantial" part of Ecodatabase? LQI will have copied much or all of the Ecodatabase data, but the amount of data LQI puts to any specific use may be much more limited. The statute does not prohibit mere copying of a substantial part of the data.

While the data was not generated or gathered by EDI, can we say that Ecodatabase is "maintained" by EDI? Under the statute, "To ‘maintain’ a database means to update, validate, or supplement the information contained in the database." We may assume that EDI keeps the power to its computer turned on and that it updates or corrects Ecoprogram when problems arise. It may also engage in some structural maintenance on Ecodatabase, e.g., changing the physical methods of storage to respond to new technologies, without changing the fields or other aspects its clients have gotten used to. None of these activities, however, fits comfortably with updating, validating, or supplementing the information contained in the database. So, EDI’s claim may be vulnerable on this ground as well.

LQI knows that it does not have EDI’s permission, having expressly turned down EDI’s offer of a license, although an adverse decision in the open-records suit might be deemed to be "authorization" from EDI (via a court order). We may also assume that EDI’s maintenance of the database through Ecoprogram, if it gives EDI database rights at all, involves a substantial expenditure of financial resources or time, although a fuller factual development might cast doubt on this assumption.

If EDI has rights under the statute, LQI could be liable for the remedies provided by the statute if (a) LQI’s making the information available "occurs in a time sensitive manner and inflicts injury on the database or a product or service offering access to multiple databases" and (b) "the ability of other parties [i.e., LQI] to free ride on the efforts of [EDI] would so reduce the incentive to produce the product or service that its existence or quality would be substantially threatened."

The statute further provides that "In determining whether an unauthorized making available in commerce occurs in a time sensitive manner, the court shall consider the temporal value of the information in the database, within the context of the industry sector involved." As we discussed in class, "temporal value of the information" seems to imply information whose economic or other value decreases solely due to the passage of time, such as information affecting stock prices or sports scores. The data contained in Ecodatabase are factual – types of equipment used, types and amounts of pollutants, etc. Whatever value they have, they do not inherently lose that value simply due to the passage of time. That is, if a plant is producing X amount of, say, sulphur dioxide per unit time, the regulators will assume the plant will continue to produce that amount in the future under similar operating conditions, unless the plant itself is modified. Indeed, there would be very little value to collecting this data if it would go stale in a matter of hours or days. Here, the information will stay in the EDI computer until it is changed by the state regulators, so it is very difficult to see time sensitivity as a factor in the analysis.

Moreover, the statute defines "inflicts an injury" to mean "serving as a functional equivalent in the same market as the database in a manner that causes the displacement, or the disruption of the sources, of sales, licenses, advertising, or other revenue." LQI’s use of the information does not compete with EDI’s primary market at all. The only market LQI’s use would disrupt is the market for the raw data licensed to companies like LQI, but this cannot be part of the definition of injury under the statute, because it would mean that every use was an injury: Whoever is using the data in a different product automatically comprises a "market" for the data. None of LQI’s proposed uses of the data can or will displace EDI’s revenue sources in the state environmental regulatory agencies.

Finally, it is difficult to see how LQI’s "free riding" can undercut incentives for EDI to create programs like Ecoprogram and maintain something like Ecodatabase. EDI got into this business because it saw a market in the state environmental regulators. It still has that market, notwithstanding that LQI is using data from the EDI database for purposes other than that for which the database was designed. Indeed, that use may actually increase demand for the Ecofacts system, because LQI’s mission in life seems to be to increase environmental regulation, which may make the need for a product like Ecofacts even greater.

The statute also contains an exemption for nonprofit institutions pursuing nonprofit educational or scientific purposes, if the court determines that the use is reasonable. LQI might be able to qualify under this exemption if it otherwise were liable (which it is not), but we would have to investigate its use more carefully. In particular, some of its proposed use seems clearly to be "educational," but other aspects may be "political," and there is an interesting issue of whether a political purpose will qualify. The case against liability under the statute is so strong without recourse to the exemption that we will not dwell on this question.

The statute also contains an exclusion for databases "generated, gathered, organized, or maintained by a Federal, State, or local governmental entity, or by an employee or agent of such an entity, acting within the scope of such employment or agency." The disjunctive "or" in this exclusion may be of crucial importance. Ecodatabase, as discussed above, is maintained (and organized) by EDI, a private company. However, the information contained in Ecodatabase is generated and gathered by state officials. The government has essentially paid for the collection of this data, and the fact that it licenses EDI to "maintain" the data (if that is what EDI does within the meaning of the statute) does should not convert the data to private property. That appears to be the intent of the exclusion. This exclusion should therefore apply as well, reinforcing LQI’s claim that it is entitled to the information under the state’s open records law.

The DMCA, in § 1201(a)(1), prohibits circumvention of technological measures governing access to works protected under title 17. If we assume that EDI has a copyright in its compilation, via its selection of categories, we have a work protected under title 17. Moreover, if we assume that EDI has rights under H.R. 3261, which (as given in the facts) is codified in title 17, we have another work protected by that title. The facts tell us that Ecoprogram is designed to make it difficult to print out or electronically copy Ecodatabase in its entirety but that a skilled programmer with access to source code could do it easily. Is Ecoprogram a technological measure that effectively controls access to a protected work (i.e., Ecodatabase)? Section 1201(a)(3)(B) says a measure effectively controls access if, in the ordinary course of its operation, it requires the application of information or a process or treatment to gain access to the work. Is that the case here?

Under the facts as given, anyone, whether skilled or unskilled at computer programming, does have easy access to Ecodatabase if he or she also has Ecoprogram. It would not be all that easy to copy the database in its entirety, because such a copier would have to access the data one datum at a time, which we are told would be both tedious and time-consuming. But the access control definition does not look to how hard or easy it is to read or use the work once access has been made. It looks to whether in the ordinary course of operation the user has to apply information or a process, with the authority of the copyright owner, to gain access. Here, the user of Ecoprogram needs no further authority of the copyright owner to gain access, tedious though it may be to transfer all of the stored data onto paper or into a different computer. I conclude that EDI has not employed a technological measure that effectively controls access to Ecodatabase, which precludes liability for violating § 1201(a)(1).

Another way to look at this would be to say that the design of Ecoprogram to make it difficult to print out the entire database is not a 2101(a) access-control measure but rather a 2101(b) copy-control measure (assuming Ecodatabase, or part of it, is protected under title 17). But 1201(b) does not prohibit direct acts of circumvention of copy-control measures; it only forbids trafficking in devices that do so, leaving to copyright law (title 17) the job of remedying any actual circumventions.

If I am wrong in this conclusion, there still may be an issue of whether the technological measure is "effective" in controlling access. The argument in Reimerdes that, because Jon Johansen managed to break the CSS encryption system, the system was ineffective was rejected by the courts, and subsequent courts have emphasized the language that the measure is effective if "in the ordinary course of its operation" it requires the application of information or a process. Here, the ordinary course of the operation of Ecoprogram is to prevent rapid copying or printing of the data. To achieve rapid copying, one must reverse engineer the object code of Ecoprogram and figure out how to modify the code to achieve this goal. One might say that this is a "process" or "application of information" without the authority of the copyright owner. But this argument applies to any modification of the functions of a computer program, which is itself a copyright-protected work. In other words, any computer program controls access to its own code by requiring decompilation and analysis in order to change its functionality. Section 117 of the Copyright Act explicitly envisions that owners of lawful copies of computer programs can make modifications necessary to achieve their desired functionality, and although we did not study this in detail this semester, courts have held this to mean the functionality desired by the user, not the functionality as provided by the copyright owner. Consequently, Reimerdes may not be controlling on these facts. The copyright owner here added no special device to the system to make it hard to copy. Rather, EDI simply supplied a program with limited functionality in the hope that fewer copies of the database were made. I would conclude that Ecoprogram is not an effective technological measure for controlling access.

The reverse engineering exemption under § 1201(f) does not appear to be directly relevant to the facts of this case. It allows circumvention of access controls on computer programs for the purpose of achieving interoperability with other programs. Here, the access control (if there is one) is on Ecodatabase, not Ecoprogram. Moreover, LQI is not seeking to obtain interoperability information between an independently created program and other programs. It just wants the unprotected information itself, for its own uses. In addition, no negative implication should be taken from § 1201(f) for the reverse engineering activity actually engaged in by LQI (i.e., the argument that if reverse engineering is not permitted by 1201(f) it is not permitted at all), because § 1201(c)(1) states that the section does not affect rights, etc. under title 17, including fair use. LQI’s reverse engineering activity was undertaken as a fair use.

Another provision that we did not discuss in class that may be applicable here is 1201(d), which a nonprofit library, archives, or educational institution to gain access to a commercially exploited work (Ecodatabase) solely to make a determination of whether to acquire a copy of the work for the sole purpose of engaging in noninfringing conduct. There is at least an issue of whether LQI is a "nonprofit library, archives, or educational institution." Moreover, LQI’s activities with respect to Ecofacts goes beyond simply determining whether to acquire a copy. LQI does not want a copy of Ecodatabase at all except possibly as an intermediate step to getting its factual content. This exemption seems aimed at breaking access controls just to see what the work is, and LQI already knows that. Probably this provision will not add to LQI’s case under the DMCA.

The tort of trespass to chattels traditionally required intermeddling with the chattel that causes actual harm to it. In eBay v. Bidder’s Edge the court broadened the traditional view of trespass to chattels by looking not to actual harm to the chattel itself (eBay’s computer system) but to the potential harm that the court thought would result if the activity were to go unchecked. The intermeddling with the chattel was the unauthorized sending of electronic signals to eBay’s computer to obtain information, precisely the activity engaged in here by LQI. Presumably the same reasoning on damages would apply here, too: If hundreds of LQI wannabes were to gain access to Ecofacts and download the database, conceivably it would place a small strain on on EDI’s computer. But even that is more of a stretch than in Bidder’s Edge, where the robot visited the eBay site some 100,000 times per day. LQI’s "interference" with the EDI computer is much more limited than that.

Moreover, the California court in Intel v. Hamidi reaffirmed the traditional view of the trespass to chattels tort by requiring proof of actual damage to the chattel, even for the purpose of an injunction. Here there is no evidence in the facts to suggest that LQI’s extraction of the information contained in Ecodatabase caused any hard to EDI’s computer. Therefore, the absence of damage should preclude any claim by EDI under the trespass to chattels doctrine.

Section 1030(a)(2) provides "Whoever intentionally accesses a computer without authorization or exceeds authorized access, and thereby obtains . . . information from any protected computer if the conduct involved an interstate or foreign communication . . . shall be punished as provided in subsection (c) of this section." Section 1030(a)(5) provides "Whoever intentionally accesses a protected computer without authorization, and as a result of such conduct, causes damage [exceeding $5,000 or personal injury or to a computer system]" is similarly punishable. A "protected computer" means any computer that is used in interstate commerce, i.e., any computer hooked up to the internet. These are criminal provisions, but 1030(g) allows damages remedies exceeding $5,000 and injunctions as well.

It is difficult to see a violation of 1030(a)(5) here because of the absence of damages. However, 1030(a)(2), if read literally, covers any unauthorized access to a computer that supplies information. Here, LQI has taken "information" from Ecodatabase, and Ecodatabase is stored in a computer that is hooked up to the internet. Its access was intentional, and it knew that it did not have authorization, at least from EDI. LQI would seem to be in the soup at least for potential criminal liability. Maybe winning the state-law open-records suit will supply the necessary "authorization." This statute was designed to help prosecute hackers, but we saw in class that its broad provisions, especially 1030(a)(2), can be and have been used in other disputes, such as the trespass to chattel cases.