researching. One of the outstanding results of this project was the creation of a names identity database—a prosopography—that helps scholars sort out the identities of the formal and personal names that appear throughout the letters and other documents in the archives. After a decade of use, the data-entry system now needs to be upgraded, and MAP is using the occasion also to reconceive its fellowship programs. It will continue to have a small number of residential fellows, but is now planning for them to be of shorter duration so that it can also establish long-distance fellowships for individuals as well as a program for distance learning. For both of these new initiatives, MAP would digitize relevant files for the research or course topic but then still require the fellows and the students under the supervision of the course instructor to catalog at a distance these files as part of the interaction. Another example is the work of Greg Crane, the classicist at Tufts, who established the Perseus database and has lately been hugely imaginative and productive in thinking about “What to do with a Million Books.” He and his research team have selected a corpus of books from the classical canon, worked with the University of Toronto and other libraries to ensure that these works find their way into the work flow of the mass digitization projects. His team then obtained library assistance and created a fully faceted, master bibliography of these selected works.25 This initiative demonstrates that one solution to the metadata problems that are rampant in Google books might be to distribute the effort to self-organizing scholarly teams that care about specific parts of the corpus and will invest the necessary effort to correct and make it usable for scholarly purposes. Crane and his team are also working with information specialists to engage other scholars and their students in developing and implementing the methods for applying linguistic markup to the corpus to facilitate machine analysis. Crane’s efforts seem to me to provide a model that could easily be emulated by other scholarly teams in other fields. Connecting Collections This brings me to the third area of development I would ask you to consider: Can we develop new and reliable methods to link related special collections across institutions? We have been exploring this area at the Mellon Foundation RLI 267 38 The Changing Role of Special Collections in Scholarly Communications ( C O N T I N U E D ) DECEMBER 2009 RESEARCH LIBRARY ISSUES: A BIMONTHLY REPORT FROM ARL, CNI, AND SPARC …one solution to the metadata problems that are rampant in Google books might be to distribute the effort to self-organizing scholarly teams that care about specific parts of the corpus and will invest the necessary effort to correct and make it usable for scholarly purposes.