The MIT libraries were called upon to recommend a metadata scheme for the resources contained in MIT’s OpenCourseWare (OCW) project. The resources in OCW needed descriptive, structural, and technical metadata. The SCORM standard, which uses IEEE Learning Object Metadata for its descriptive standard, was selected for its focus on educational objects. However, it was clear that the Libraries would need to recommend how the standard would be applied and adapted to accommodate needs that were not addressed in the standard’s specifications. The newly formed MIT Libraries Metadata Unit adapted established practices from AACR2 and MARC traditions when facing situations in which there were no precedents to follow.

This paper describes work in progress on the research project CWSpace, sponsored by the MIT and Microsoft Research iCampus program, to investigate the metadata standards and protocols required to archive the course materials found in MIT's OpenCourseWare (OCW) into MIT's institutional repository DSpace. The project goal is "to harvest and digitally archive OCW learning objects, and make them available to learning management systems by using Web Services interfaces on top of DSpace." The larger vision is one of complex digital objects (CDOs) successfully interoperating amongst MIT's various learning management systems and learning object repositories, providing archival preservation and persistent identifiers for educational materials, as well as providing the means to richer shared discovery and dissemination mechanisms for those materials. The paper describes work to date on the analysis of the content packaging metadata standards METS (Metadata Encoding and Transmission Standard) and especially IMS-CP (IMS Global Learning Consortium, Content Packaging), and issues faced in the development and use of profiles, extensions, and external schema for these standards. Also addressed are the anticipated issues in the preparation of transformations from one standard to another, noting the importance of well-defined profiles to making that feasible. The paper also briefly touches on the DSpace development work that will be undertaken to provide new import and export functionalities, as the technical specifications for these will largely be determined by the packaging metadata profiles that are developed. Note that the degree of interoperability considered herein might be referred to as "first level," as this paper addresses the packaging metadata only, which in turn is the carrier or envelope for the descriptive (and other kinds of) metadata. It will no doubt be an even more challenging task to ensure interoperability at what might be referred to as the "second level," that of semantic metadata.