Two Cultures: A Social History of the Distributed Library Initiative at MIT

Nina Davis-Millis and Thomas Owens

Before the 1990s, libraries and campus computing centers had a long history of working together, but the typical relationship could hardly be called a partnership. Libraries prevailed on the computing centers to provide expertise and services but generally operated their own stand-alone minicomputers. Housed in library buildings and operated by library staff, those systems delivered information solely to library locations. Dial-up access represented the furthest reach of the library computing environment.

Several trends during the latter half of the 1980s, however, led to a nationwide reexamination of this precursory relationship. On the computing side, the microcomputer revolution changed the nature and scope of the services required by traditional customers. Accounting and inventory applications moved to the desktop; service priorities shifted from centralized storage and access to communications and integration. A first step toward meeting these priorities was a campuswide network and, by 1990, most universities either had such a network or had plans for one. The Internet created the opportunity for electronic mail, the “enabling application” that further pushed demand for network access.

Library automation also reached a critical juncture at the turn of this decade. Many institutions had automated during the integrated-systems boom of the early 1980s, and those first-generation systems were now due for replacement. Initially wary of computers, librarians now had a taste for them and wanted to supply automated access to a wider range of data, especially journal citation data. The CD-ROM revolution, although generally taking place on local area networks, clearly pointed to an almost unlimited range of automated information resources. Most important, the accelerating growth of campus networks linked to the Internet led to ever-increasing user demands for remote access.

In short, by 1990 or so, libraries needed to expand their services in ways that required deeper technological expertise than they possessed. Computing centers, for their part, were freed from traditional tasks by the microcomputer and had the time to devote to library needs. They also operated the communications infrastructure libraries wished to exploit. These trends pushed the two academic centers into each other’s arms. On some campuses, there was a merger—sometimes led by a librarian, sometimes by a systems professional.

A Brief Narrative of the Distributed Library Initiative

The recent history of MIT libraries and the campus computing center (called information systems, or IS) illustrates these trends. The libraries installed their first integrated system in 1984; by 1990, that system had reached its limits in size and functionality. It had been extraordinary when purchased, but technological change had outmoded the original proprietary hardware, which emphasized dedicated, polled terminals hostile to network communications. On the IS side, MITnet had been an early member of the Internet family and provided a firm foundation for universitywide connectivity. “Athena,” the students’ computing environment, for example, was a famously successful project providing distributed services to more than 1,500 Unix workstations.

However typical MIT was, as far as we can learn, no one ever seriously considered merging IS with the libraries. Actually, we sought a marriage rather than a merger—a real partnership that would allow each organization to bring its specific expertise to the problem of networked information. We wanted this partnership to grow simultaneously and organically at all staff levels—to foster a convergent understanding of goals, values, and methods on both sides.

Planning for the partnership started in 1990 at the highest administrative levels, with Marilyn McMillan, IS’s director of information systems planning, and Greg Anderson, the libraries’ associate director for systems and planning, meeting to discuss their organizations’ mutual interests and goals and later arranging for larger meetings with other administrators in IS and the libraries. Those meetings led to a basic formulation of the respective strengths of the two organizations: The libraries determine content; IS provides infrastructure. On one side, the libraries were organizationally and intellectually better suited for judging the content of networked information—they had established methods for keeping current with the curriculum, for monitoring trends in campus research, and for understanding the goals of scholarly communication. On the other side, IS was organizationally and intellectually prepared to provide the expertise and tools required for operating and maintaining the infrastructure to deliver the information the libraries selected. Moreover, IS could integrate the delivery of this information into the MIT computing environment.

It became clear, however, that the provision of substantial amounts of networked information would require more effort than either organization was then prepared to supply on its own. At that point, Jim Bruce, vice president for information systems, and Jay Lucker, director of the MIT libraries, realized that ongoing strategic contact was required to make the partnership critical to both organizations. This led to the establishment of the first formal policy-making collaboration, The Libraries/Information Systems Steering Group, which soon became known by its e-mail address, LIBTALK. It set out to create a strategy for providing electronic library services to the MIT community—and the world—that developed into a project known as the Distributed Library Initiative (DLI). LIBTALK agreed that the partnership would be both long term and central to both organizations. It would be guided—but not driven—by specific, current projects. LIBTALK then initiated a process of mutual education at all staff levels.

Technical staff from both organizations met for the first time in 1991. Libraries staff presented a history of library automation, a brief introduction to the MARC record structure, and an explanation of the concept of authorities. IS staff explained network services, the Athena computing environment, and security protocol. These meetings continue to this day but in a different format. Now they are lunchtime talks with open attendance and both organizations urge all staff to attend. Presentations have ranged from a detailed explanation of the Z39.50 protocol, to a discussion of copyright issues in the electronic forum, to talk about current projects. (The technical group became known, as LIBTALK had, by its e-mail name—ELIBDEV—for Electronic Library Developers.)

To foster working relationships, IS and the libraries initially worked on a few small projects. OWL (On-line with Libraries), for example, was a jointly developed, Athena-based online reference service. At about the same time, the “library menus” appeared on Athena. These pull-down menus led to a variety of information services available either online or in the libraries’ paper collections. On a broader level, we began examining infrastructure issues. The Z39.50 protocol was key to our plans and, in 1991, representatives from IS and the libraries joined the Z39.50 Implementors Group, an international forum of developers for the nascent Z39.50 protocol.

As such small projects proliferated, the partnership saw the need to coordinate the various activities. A second formal group was created (again named for its email address—DLICC—for Distributed Library Initiative Coordinating Committee). In addition, the Laboratory for Computer Science research group Library 2000, headed by Professor Jerome Saltzer, was keenly interested in electronic libraries, and members of this group joined the coordinating committee. To help with copyright issues, the Intellectual Property Counsel became a regular attendee.

For major projects, the membership of the coordinating committee is now quite flexible, though a core group remains. As the two organizations worked together and learned more about each other, they sought out users to advise them on the creation of new services. For example, in 1992, the libraries and IS joined with the Committee on Academic Computing to present a DLI Day, which attracted more than seventy participants from across the university, including staff, researchers, faculty, and students. Small groups focused on specific information needs, opportunities, and constraints.

All this activity led to the eventual development of the DLI vision by the IS/Libraries Steering Group in 1993. This vision statement concluded with a vivid forecast of the DLI five years hence, for 1998:

The way MIT and its people do research and pursue education will be revolutionized by enriched access to all forms of information at their fingertips. Sitting at a workstation in the classroom or laboratory, in the sorority, or in the airplane, anyone can retrieve, manipulate, interpret, and integrate information into their personal knowledge banks. They can easily move among personal, on-campus, and worldwide resources to find, evaluate, sort, and store information. MIT students, researchers, educators, and administrators, freed from the drudgery of information management, are now better able to work together in putting that information to use in the advancement of humanity.

With this vision as a guide, interest in access to networked information spread to the creation of still another formal IS/libraries group, the Network User Team (NUT). NUT’s IS members include a programmer, a faculty liaison, and the manager of its Educational Planning and Support unit; NUT’s libraries members include representatives of public services, collections, technical services, and library systems. This group manages new releases of existent applications, develops access to new resources, and coordinates technical support of applications and hardware with functional support of content and training. NUT’s projects have included adapting each successive release of WILLOW, a Z39.50 client developed at the University of Washington, for use in making databases available over Athena; evaluating available interfaces for the online Oxford English Dictionary and implementing a Web version employing Open Text software; and providing ongoing support of networked information resources. Support to software is provided by the Library Systems Office and IS; questions regarding intellectual content or training are handled by reference librarians and Athena faculty liaisons.

Currently, the steering group, the coordinating committee, and NUT are the only formal joint IS–Libraries groups. However, nearly every new project creates ad hoc groups on varying levels of formality. Recently, we have experimented with software development by open invitation: Once a project is defined and a core team identified, meetings can be attended by any interested staff and the product goes through extensive testing and review. The core team formally answers each suggestion, explaining why each one was or was not adopted.

At about the same time (1993) that the steering group developed the DLI vision, it issued a general schedule of network development. Since then, DLI has been mostly on track. Working with the University of Washington and Stanford, for example, MIT created an early Z39.50 client (WILLOW) with code written by both IS (underlying protocol drivers) and library staff (MARC record management). The library Web soon superseded the library Gopher. We mounted our first local databases using BRS software, then migrated to OCLC’s Newton; software was written to provide automated access to OCLC’s FirstSearch databases through Athena; and a team of librarians and IS staff worked on Elsevier’s TULIP project (delivering images of that publisher’s engineering journals to the desktop using both WILLOW and the Web).

DLI: A Social History

Sustaining our drive through the next few years is crucial to DLI’s viability in the long run. At this point, it is fair to say that IS and the libraries have developed a fruitful, mutually respectful working relationship. From the beginning, both IS and the libraries consciously acknowledged the cultural differences of the two organizations while focusing on common goals. Although we always emphasized each other’s professional expertise and distinct areas of competence, we also fostered “technology transfer” between the two organizations. Our plan was not to have one group adopt the culture and methodologies of the other but, rather, to let a third culture evolve out of shared philosophies and objectives.

Indeed, we consider that the awareness of cultural differences between the two constituencies, together with the decision to invest staff time in exploring them (as well as professional similarities), has been a key element in the success of the partnership. Another salient factor has been the active participation of upper-level administrators from both units in setting goals and priorities, concurrent with the participation of frontline staff in the actual design and delivery of services. These considerations, both formal and informal, have enabled groups in each organization to share their respective expertise, skills, and values.

A brief review of the literature on collaboration between academic libraries and campus computing organizations provides a broader perspective on how typical our experiences may be. In this literature, a great deal of analysis has been done on the pros and cons of merging such units. However, relatively little attention has been paid to the benefits of sharing cultures, or even to descriptions of the respective professional cultures.

David C. Weber points out that the library profession has a long shared history of standards, philosophies, and procedures whereas campus computing has a strong technical orientation characterized by a volatile entrepreneurial environment and rapidly changing technology.1

Sheila D. Creth, in a survey of discussions among library and computing administrators and managers, observes that campus computing staff come from a variety of backgrounds:

Since they do not share a common professional and academic preparation, there is no socialization process for a computer professional prior to accepting a position in a computer organization. They considered this lack of shared professional philosophy and values to be a major contributor to an ethos in which individual action and thought is more highly regarded than a focus on the views, standards or values of the collective group. In contrast, this group of [computing] professionals considered that librarians with their similar educational backgrounds experience a process of acculturation in which they develop a shared philosophy and common values. Therefore, they are more likely to act within the boundaries of accepted professional beliefs and behavior, being less likely to act independently.2

Creth’s survey also dealt with the respective values of the two professions. Computing professionals were regarded as sharing a technical orientation, as encouraging entrepreneurial behavior, and as valuing creativity, whereas librarians were seen as having a service orientation, a consensual approach, and fiscal responsibility. Both groups were identified with a professional orientation of being focused on the world of information and the consumers of information, and of being concerned with their own institution (and hence, one may assume, with their local customers).3

As Kristin McDonough points out, we should distinguish “subcultures” and “vested interests” within and between both the library and the computing professions:

There used to be talk of separate ‘cultures’ in our own profession, as exemplified by public services and technical services librarians. Today, technological advances have forced a crossover. . . . Another salient point is the competition that has existed on some campuses between administrative and academic computing. Rather than ‘cultures,’ might not the issue be the pull between vested interests?4

To expand our perspective of MIT’s merger efforts, we interviewed McMillan and Anderson about the early stages of libraries–IS collaboration and later invited the participants of various joint efforts to a focus group. Overall, our colleagues confirmed the sense that MIT’s experience corroborates many of the observations chronicled in the literature.

Looking back on the early stages of collaboration, Anderson and McMillan remembered that each organization expected to benefit from building a partnership with the other. Although both organizations hoped to be the primary provider of electronic library services over the network, IS had the technical expertise and the hardware as opposed to the libraries’ well-established connections to instruction, scholarship, and MIT’s academic life in general.

Ironically, each unit saw the other as being at an advantage politically. The libraries, were concerned that they would be absorbed and controlled by IS which, in turn, feared that the libraries had deeper service contacts across campus and could build powerful alliances demanding unsustainable services. Picking up on the “marriage” metaphor, Anderson and McMillan agreed that their organizations saw the possibility that the partnership would be a “deadly embrace” as each tried to leverage resources from the other while retaining complete control—ending with both parties squandering staff and financial capital on fruitless projects. To avoid that, they recognized that some kind of collaborative model would have to be developed. As they searched for the right model, they kept the cultural differences constantly in mind.

The initial focus was actually on defining the scope of the partnership; awareness of each other’s organizational structure, and then the respective professional cultures, followed as the collaboration took shape. McMillan and Anderson arranged high-level meetings that avoided discussions about specific projects but, instead, focused on sharing organizational goals and values: What does your organization wish to accomplish, and what is important to you? The meetings were carefully designed not to raise expectations or to preselect outcomes. Clearly, McDonough’s vision of a possible “pull between vested interests” was perceived in the formative stages of the libraries–IS partnership. The great extent to which this pull was overcome can be attributed to all parties having taken the time necessary to accommodate one another’s organizational frameworks and to explain decisions in one another’s terms. With the top management in the two organizations being committed to collaborative dialogue, the front-line staff who worked on specific products experienced a rich, creative tension rather than a clash of cultures.

The focus group we organized for this study was composed of libraries and IS staff, all of whom had been active in at least one collaborative project. We asked participants to comment on the two professional cultures of librarianship and computing. They perceived librarians as tending to be orderly, striving for perfection, and having a sense of responsibility to the public. IT professionals were regarded as being eager to make resources available—”willing to throw stuff out there,” as one participant phrased it—but also inclined to assume that almost anything is subject to change. Whereas librarians were seen to believe that things should stay as they are (and this may explain their urge to “strive for perfection”), IT culture holds that change will probably be beneficial (an assumption that can backfire if customers are already satisfied with a service or product).

Despite that fundamental difference in outlook, the librarians valued certain attributes of their IS colleagues: flexibility and a positive orientation to change, combined with a task orientation and a focus on results. IS participants, in turn, appreciated librarians’ respect for the past and their ability to apply lessons or knowledge from the past to the future. IS participants generally seemed to have gained from joint projects an appreciation of the depth of professional librarianship.

A group whose experience was distinctly more ambivalent was library systems personnel. This unit’s dual role seems to have created ambiguity and frustration for its members—a sense of being caught between two cultures, each with a clearly defined area of expertise. (In future research, we hope to explore the cultural and professional dynamics of library computing units in academic libraries during the Internet era now under way.)

Denis H. J. Caro and Amarjit S. Sethi contend that of the factors that shape a technology plan or project, the social ones are more important than operational, technical, or economic ones.5 McDonough agrees that “among the daunting challenges that lie before campus information providers, the human issue of cooperation is as important—and as thorny—as the technological ones of connectivity and networking.”6 MIT libraries and IS have avoided many of the notorious social-interaction pitfalls by acknowledging and fostering the human side of collaboration; by giving staff members throughout our respective hierarchies a chance to participate in planning and information sharing, both in formal and informal settings; and by encouraging the building of partnerships that are rich in personal satisfaction and not limited to tapping the knowledge of the individuals involved.

Nevertheless, this approach had some initial costs. McMillan and Anderson spent three to four hours preparing for every hour-long meeting. They agree that focusing only on projects would have supplied more successes in the beginning. But it is by no means clear that quick successes would have fostered the deep working relationships the two organizations have achieved. These relationships have moved beyond the DLI into other projects, some removed from information retrieval. For example, IS invited library staff to join the task force charged with creating and maintaining software development methods at MIT. In addition to creating procedures for identifying and tracking requirements, the team examined software development tools and project planners. Library staff were asked to advise the reengineering of the IS help desk, and IS also chose a librarian to be the CWIS administrator. On the libraries side, IS staff now run all the major servers used by the libraries; they routinely offer security audits of library-operated systems. The libraries also make sure that all licenses go through the IS Intellectual Property Counsel.


Traditionally, campus computing centers and libraries have had different “cultures” that emphasize rather distinct values and goals. MIT elected to build a long-term collaboration, a partnership that would require ongoing relationships at all levels of the two organizations. To accomplish this, it focused on brief, project-oriented tasks while building a work environment that directly addressed the issues raised by cultural differences. This approach, characterized by concurrent investments in specific projects and underlying working relationships, has resulted in an effective and productive partnership, highly valued by members of both organizations. Participants in our joint ventures have profited by a deeper understanding of the professional values of their counterparts, and the resulting services have benefited from respective strengths of our two professions.


  1. David C. Weber, “University Libraries and Campus Information Technology Organizations: Who Is in Charge Here?” Journal of Library Administration 9, no. 4 (1988): 5–19.
  2. Sheila D. Creth, “Creating a Virtual Information Organization: Collaborative Relationships between Libraries and Computing Centers,” Journal of Library Administration 19, no. 3/4 (1993): 120–21.
  3. Ibid.
  4. Kristin McDonough, “Prior Consent: Not-So-Strange Bedfellows Plan Library/Computing Partnership,” ERIC/ED 364225 (Apr. 12, 1992), 10–11.
  5. Denis H. J. Caro and Amarjit S. Sethi, “Technology Strategy: The Role of Strategic Planning and Monitoring Systems,” in Strategic Management of Technostress in an Information Society, ed. Amarjit S. Sethi et al., (Lewiston, N.Y.: C. J. Hogrefe, 1987), 35.
  6. McDonough, “Prior Consent,” 4.