Development of an Operational Library

An emerging and important aspect of the ADL project has been to develop a full-service operational digital library from the testbed. We believe that this is a critical endeavor. From the point of view of a DLI project, the most important reason for developing an operational facility is that it is not possible to carry out useful research in many areas of digital library technology without a testbed that is essentially an operational library. It is not possible, for example, to investigate many issues relating to users' response to library functionality in general and to its user interface in particular if the library does not support useful collections and library services in a user-friendly manner. This, however, appears to be a central aspect of the definition of an operational library facility. A second reason for constructing an operational facility is that there are many users of geospatial information who would benefit from such facility.

We are employing two approaches in moving towards an operational library facility on the basis of our current and emerging testbed technology, with one approach being at the UC Systemwide level, and the other approach at the local level. This year, the UC systemwide process has proceeded along two paths (1) working with the various UC map and government publications librarians and (2) working directly with the Office of the President of UC and the Director of Library Automation.

In relation to the first path, a UC/Stanford working group of map and government publication librarians was created for development of an implementation plan to use ADL for distributing spatially indexed information within University of California. The processes for migrating traditional library functions to an operational digital library environment are not well documented, thus, this group was formed to investigate how this might be accomplished. In June 1996, an invitational meeting was held at UCSB for the development of an intra-campus working group and for constructing an operational plan to extend the Alexandria testbed to the UC/Stanford community as a whole. ADL's emphasis on distributed spatially-indexable data management and delivery via the Internet was the focus of the meeting.

The ADL/Library team felt it was time to invite our colleagues to participate in the testbed; due in part to having a working ADL web prototype available for testing. We felt that planning for an operational system must include experiences and expertise from many archives, disciplines, and business-related environments. Starting this process within UC seemed a logical action. Those attending from each campus, (UCB, UCD, UCLA, UCSD, UCI, UCSFO, Stanford and UCSB), were associate university librarians, map librarians, and government document librarians. Presentations of ADL technology were made and demonstrations given of the prototype testbed. The group consensus was to actively move forward by preparing a document that would, over time, evolve into an implementation plan. Using ADL as a service component of the University of California library system, and perhaps later, the State of California, was envisioned. A small committee was selected to construct the first draft for general group review.

A mission statement was formulated that would capitalize on ADL's research to build a UC-Wide operational system for searching and accessing distributed spatially-indexed data and provide tools for data discovery, data management, and user services via the Internet. The document was to outline processes for the development of:

1. a multi-campus generalized operational model for a digital library;
2. a multi-Campus ADL operational requirement;
3. prototype scanning projects on each campus that would support the ADL testbed ( e.g. geospatial datasets and/or georeferenced data collections to put on line (see section on collections);
4. the identification of individuals or academic areas that might be interested in working on such a project; it would include those interested in interface issues, catalog development, user services, interaction between ADL and non-UC agencies who create or archive spatially index information.

At the local level, a UCSB Library Operational Library Group has been established. Its mission is to develop an end-to-end model of what an analog/digital operational library might look like. This includes reviewing all current library operations, from technical processing to user services, and then suggesting a migration path towards linking the best of both. Research issues such as how to determine content, measure success, and model operational differences between a digital library and current library services are part of the charge of this group.

The team is charged, over the next year, with better understanding the operations of an analog library and how they should be linked to ADL. The group will determine, for example, related existing library data management components with proposed ADL services. In particular, ADL might hope to link its operations to existing library book-accounting services, rather than to build it own. Once the team has identified, at a fairly high level, what components complement each other (current library operations and those of an operational ADL), it will model an end-to-end set of components. Library components being examined are:

1. Archive management (e.g. personnel structures, systems, budget);
2. Collection management (e.g. data selection and acquisition);
3. Technical Processing (e.g. cataloging, digitizing, and loading data into the store);
4. Technical Services (e.g. user support for searching, browsing, and data serving (check-out, download, format, preprocess, ...);
5. User Evaluation (e.g. quality of: collections, system services, ... ).

The committee agenda over the next year is to construct a framework for building an operational digital library. Such a framework will include:

1. describing library components and procedures;
2. outlining an end-to-end model detailing sub-component operations such as the terabyte database system being constructed in partnership with Microsoft; (see below);
3. developing generic scope statements for each subcomponents;
4. constructing a visual representation of the library model;
5. identifying procedures ( from general to specific ) required for each operation within the model;
6. identifying resources required for an operational digital library, including such infrastructure resources as space, network connections, communications, power, and personnel, as well as administrative support, business processes, systems, networking, and user services;
7. identifying systems including hardware, processing and user workstations, storage (local and remote), and software;
8. identifying requirements for ingesting materials e.g. scanning, data conversion or formatting;
9. specifying an operational digital library in terms of a ``cookbook'' for other libraries;
10. determining the types of user services and support that would be required, such as a ``1-800-archive'' service.