Abstract:
Digital texts with access and use protections form a unique and fast growing collection of materials. Growing equally quickly is the development of text and data mining algorithms that process large text-based collections for purposes of exploring the content computationally. There is a strong need for research to establish the foundations for secure computational and data technologies that can ensure a non-consumptive environment for use-protected texts such as the copyrighted works in the HathiTrust Digital Library.
Developing a secure computation and data environment for non-consumptive research for the HathiTrust Research Center is funded through a grant from the Alfred P. Sloan Foundation. In this research, researchers at HTRC and the University of Michigan are developing a “data capsule framework” that is founded on a principle of “trust but verify”. The project has resulted in a novel experimental framework that permits analytical investigation of a corpus but prohibits data from leaving the capsule. The HTRC Data Capsule is both a system architecture and set of policies that enable computational investigation over the protected content of the HT digital repository that is carried out and controlled directly by a researcher.