Effective information integration and reutilization : solutions to technological deficiency and legal uncertainty

Abstract

The amount of electronically accessible information has been growing exponentially. How to effectively use this information has become a significant challenge. A post 9/11 study indicated that the deficiency of semantic interoperability technology hindered the ability to integrate information from disparate sources in a meaningful and timely fashion to allow for preventive precautions. Meanwhile, organizations that provided useful services by combining and reusing information from publicly accessible sources have been legally challenged. The Database Directive has been introduced in the European Union and six legislative proposals have been made in the U.S. to provide legal protection for non-copyrightable database contents, but the Directive and the proposals have differing and sometimes conflicting scope and strength, which creates legal uncertainty for valued-added data reuse practices. The need for clearer data reuse policy will become more acute as information integration technology improves to make integration much easier. This Thesis takes an interdisciplinary approach to addressing both the technology and the policy challenges, identified above, in the effective use and reuse of information from disparate sources.

(cont.) The technology component builds upon the existing Context Interchange (COIN) framework for large-scale semantic interoperability. We focus on the problem of temporal semantic heterogeneity where data sources and receivers make time-varying assumptions about data semantics. A collection of time-varying assumptions are called a temporal context. We extend the existing COIN representation formalism to explicitly represent temporal contexts, and the COIN reasoning mechanism to reconcile temporal semantic heterogeneity in the presence of semantic heterogeneity of time. We also perform a systematic and analytic evaluation of the flexibility and scalability of the COIN approach. Compared with several traditional approaches, the COIN approach has much greater flexibility and scalability. For the policy component, we develop an economic model that formalizes the policy instruments in one of the latest legislative proposals in the U.S. The model allows us to identify the circumstances under which legal protection for non-copyrightable content is needed, the different conditions, and the corresponding policy choices.

(cont.) Our analysis indicates that depending on the cost level of database creation, the degree of differentiation of the reuser database, and the efficiency of policy administration, the optimal policy choice can be protecting a legal monopoly, encouraging competition via compulsory licensing, discouraging voluntary licensing, or even allowing free riding. The results provide useful insights for the formulation of a socially beneficial database protection policy.