A secure framework for information sharing and structured search of distributed research resources

Researchers collaborating together on a project are often separated by both geographical and institutional boundaries; in addition, they may want to use resources from disciplines that are both directly and indirectly related to their research. As a result, researchers need good search tools to help them find resources. Traditional search tools require a researcher to know the vocabularies of his discipline and any related disciplines to find these resources. To further complicate the search process, resources can be textual, graphical, audio, or a combination of the three. Since search engines traditionally have been engineered for a particular media, a researcher may have to use more than one search engine to find resources.;Collaborating researchers need a way to both share resources with other researchers and securely search these resources. Individuals having relationships with researchers who create these resources can learn of them by word of mouth and be granted access, if the creating researcher knows he/she can trust them. However, other researchers who are not directly involved with the creating researcher and thus unaware that these resources exist also need a way to locate them and request access; however, access to those resources is not likely to be granted to an unknown individual by the creating researcher without first establishing some level of trust with the requester.;To address these dual problems of searching and sharing resources in a secure manner, we have developed a framework for information sharing and structured search of distributed research resources. The framework associates descriptive metadata with a resource, then stores that metadata separately from the resource in a centralized repository. This enables researchers to secure their resources to the level of their satisfaction, yet promote their resources to others. The centralized metadata repository is secured, but in the event that it is breached, only metadata, not the resource, will be compromised.;In the framework, a resource is represented with a metadata set that is standard across all media types. Fields within this metadata set include concepts associated with a resource, in addition to other information such as author, resource title, and author contact. Concepts within a resource are connected via simple relationships. By storing concepts and relationships, a search can return resources with matching search criteria provided as well as resources with concepts indirectly related to the search criteria.;A prototype implementation of this framework has been developed to assess its feasibility and evaluate its mechanisms for representing and locating resources. This prototype is discussed and demonstrated.