Since 2000, the National Science Digital Library (NSDL) Core Integration team has been creating the infrastructure for a digital library of science, technology, engineering, and mathematics resources. That library now contains more than a million resources from approximately 100 collections. In this talk, Dean Krafft gives a short historical overview of the NSDL and describes the current NSDL community and participants. He then reviews the technical underpinnings of NSDL 1.0, a library built on metadata harvesting, and describes some of the challenges encountered. For the past year (2005-6), the project has been working on NSDL 2.0, a new version of the library built on the Fedora repository architecture. For the last part of the talk, Krafft describes this new library architecture and explains how it supports creating context for science resources, how it enhances the selection and use of library materials, and what these capabilities mean for the users of the NSDL.