摘要:Scientific research collaborations generate massive amounts of data that are assembled into collections, published in digital libraries, processed in data analysis pipelines, and preserved in reference collections. Policy-based data management systems minimize the amount of labor needed to manage the massive collections by automating the enforcement of management policies and the validation of assessment criteria. The goal is data management infrastructure that can be used to support all phases of the data life cycle, while minimizing the amount of labor needed to maintain the collection.