Promote interest in carrying out research on use of digitized resources. Possibilities: OCLC project, library schools, ALISE, individual libraries, vendors, scholars, publishers. Find out what research is currently in progress.

Consider general surveys and focus groups to assist in resolving some issues. Who should do?

Access not effective: emphasize training, non-mediated service model.

Work with public services to determine what is important to users. Is there resistance to access on the part of public services, the staff managing the Internet access, collection development people?

How will the records we create get people to materials?

Explore training and education issues. Librarians (especially catalogers, both professional and paraprofessional, and collection development practitioners) should take the lead. At this point, educating and raising an awareness of issues is more important than training practitioners. Need to raise knowledge level of software and hardware.

New coursework should prepare students to set up Internet nodes and/or Web servers, designing user-friendly interfaces; coursework to recreate the book, looking at the catalog as a system.

What should we be teaching to prepare for these shifts?-- Emphasize the engineering aspect as in engineering information systems.

After listing some of the diverse forms that networked information resources can take (e.g. electronic texts and journals, databases, Gopher and FTP sites, image data, multimedia, etc.) the group identified several areas for research and experimentation:

- defining and describing the intellectual content of networked information resources

- exploring new or extended record structures, including dependent structures to provide a deeper level of intellectual access. For example, the TEI header.

- determining user needs in a networked information environment

The group acknowledged there are problems with authentication, copyright, attribution, and permanence of electronic resources, and that these problems will not be quickly or easily resolved, but felt that work in the previously identified areas should move forward.

Finally the group considered a future model that included the automatic creation of a preliminary bibliographic record when an electronic object is created. These records would be made available for revision and upgrading after the object was registered (given a URN). Records could be searched and maintained in a traditional bibliographic system, with changes to the electronic object being automatically forwarded to the registry service and then to the surrogate record.