By Bill Steele

Cornell researchers have received a National Science Foundation (NSF) one-year grant of $799,085 to develop a proposed architecture for a new online library for science education. The award is part of a $13.5 million NSF initiative to create a National Science, Mathematics, Engineering and Technology Education Digital Library (NSDL) to give students wide access to a rich array of authoritative and reliable interactive materials and environments, according to Lee Zia, NSDL program officer at NSF. The library will cater to students and instructors from kindergarten through college and beyond. William Arms, Cornell professor of computer science, will lead a team developing a demonstration of a core system for the library -- the engine that will find the information and make it accessible for searching.

"Our grandiose vision is to bring together all digital information of potential value to scientific education, where you define both science and education as broadly as you can," Arms said. "Enormous volumes of scientific material are now available on the Internet but poorly organized for use in education," he explained. "Students and instructors have to know where these collections are and how to use them to get at the information. They need help in distinguishing good science from bad. Computer systems must recognize that different people have different requirements."

The NSDL project will be just one of several possible systems to draw on a new standard, known as the Open Archive Initiative, spearheaded by two other Cornell researchers, Herbert Van De Sompel, visiting assistant professor of computer science, and Carl Lagoze, digital library scientist in the Cornell Digital Library Research Group. After meetings with several other digital library researchers, Van De Sompel and Lagoze have developed a standard format in which the holders of digital collections can publish what librarians call "metadata," or data about data. Services like the NSDL will reach out across the Internet to index the metadata and lead users to the information they want. "All the search engines are indexing the visible web," Van De Sompel explained. "There's also the invisible web or the deep web, not accessible by the http [web] protocol. Our protocol could be something that allows you to index the deep web."

Arms and Carol Terrizzi, communications director for the Cornell NSDL project, will be co-principal investigators on the core system effort. Others on the team are Lagoze; Van De Sompel; John Saylor, director of the Cornell Engineering Library; Diane Hillmann, librarian and metadata specialist; Rich Marissa, senior software designer and implementer; and Dean Krafft, research associate in the Department of Computer Science. The team hopes to have a demonstration of its system ready for the Jan. 23, 2001, meeting of the Open Archive Initiative group in Washington, D.C. "We'll be doing a demonstration of how the whole thing might be done," Arms said, noting that it's most unlikely that what the Cornell group creates will be the final version. He said: "We are one of five or six projects developing the core structure. I believe we've got an exceptionally strong technical group, and I'm hoping that our ideas of the architecture will become part of the final system." NSF is targeting fall of 2002 for launch of a completed NSDL library system.