bodies does not rise to the level of an original work of authorship. Moreover, a mature copyright system usually affords protection only against wholesale copying of the original selection and arrangement underlying any eligible compilation of data. In the United States, this doctrine of weak or "thin" protection for factual works has been reinforced by First Amendment concerns, which some courts and commentators viewed as mandating broad access to the disparate facts that result from a compiler's efforts.23 When these doctrines apply, they greatly diminish the value of copyright protection even to database publishers who satisfy the eligibility criteria, because their exclusive reproduction and derivative work rights—as construed by the Supreme Court in Feist Publications, Inc. v. Rural Telephone Service Co.—will not normally prevent unauthorized extractions of disparate data for either competing or value-adding uses.24
Some federal appellate courts, however, have begun to rebel against the Feist decision and to reinstate stronger copyright protection for factual compilations and databases by subtle doctrinal manipulation. 25 Whether state or federal unfair competition laws could also provide some supplementary relief against the unauthorized copying of commercially valuable data that are not protected by trade secret or copyright laws remains an unsettled question, although such laws are sometimes invoked both here and abroad.26 In any event, this cyclical fluctuation between states of underprotection and overprotection is a characteristic trait of borderline subject matter that fits imperfectly within the classical patent and copyright paradigms, such as the contents of databases.27
Despite (or perhaps because of) the relatively weak legal infrastructure governing use of data, a thriving market for compiled information has grown up, and U.S. publishers appear to play a dominant role in it,28 although it is important to emphasize that this market has been largely concerned with nonscientific data and information. This industry seems largely characterized by niche marketers who supply and dominate specific market segments. The limited size of these segments and the relatively high startup and servicing costs seem to deter second comers from readily entering such markets.29 In other words, once the threshold level of investment has been crossed, the first comer tends to take the relevant market segment as a whole.
The public sector nonetheless has remained largely immunized from the potential abuses of market power inherent in this situation, owing both to its subsidized status and to the long-standing legal tradition that denied copyright protection to works produced by U.S. government agencies.30 As a result, data provided by federally funded projects have flowed through the domestic innovation system with few legal impediments (see Box 5.1), and legal disputes about