New Libraries System Installed; Search for Director ProgressesBy Saul Blumenthal
Associate Night Editor
The MIT libraries will undergo two important changes this summer as a search committee narrows down its choices for a new director to replace Jay K. Lucker, who is retiring, and administrators work to install a new computer operating system.
The search committee has winnowed a national field of candidates down to nine, according to Professor Peter S. Donaldson, head of the literature section and chair of the committee.
Six of the nine candidates have already given presentations to the committee on the topic "The Role of the Research Library in the 21st Century," and the rest will do so over the next two weeks, Donaldson said. While he declined to name names, Donaldson did say that none of the candidates are MIT staff or faculty.
The committee plans to report its recommendations to the provost and president within the next several weeks, Donaldson said. The committee hopes they will choose a new director by Sept. 1, a day after Lucker's Aug. 31 scheduled retirement.
Committee members noted that the new director must have a strong interest in and commitment to technology. Besides being interested in "traditional library services," the new director must "lead the MIT Libraries in terms of information technology," said Dutch D. Chung G, one of two students on the 11-member committee.
The top candidates must have "the capacity to lead change in the library world" by making use of the technological and information revolution being brought about by the Internet and the World-Wide Web, Donaldson said.
"Everyone considered has had great interest in information technology," Donaldson said. The new director "doesn't have to be a technologist, but should have familiarity with information technology and with what is best."
In general, the committee is looking for a "candidate interested in user input and improving services in the libraries," Chung said.
New system ready by fall
In addition to a new director, the libraries will be moving to a new computer operating system this summer, said Greg Anderson, MIT Libraries' associate director for systems and planning. The new system will be much more powerful, have more sophisticated indexing, searching, and processing capabilities, and be better structured than the current GEAC 8000 system, Anderson said.
The $360,000 system is the result of a five-year-long collaboration between the Libraries and Information Systems to improve public access, Anderson said. Originally scheduled for completion late last year, the project was delayed last June when the company originally chosen to design the system changed its plans ["MIT Libraries' Barton Replacement on Hold," Sept. 23].
"For its time, [the current system] served MIT well," Anderson said. "In today's computing and networking environment, however, the system is not powerful, open, or flexible enough to meet the requirements of the MIT community."
The conversion to the new system begins next week; it should be up and running by the end of the summer, Anderson said. The transition has made some services unavailable, such as placing holds or recalls on books and renewing books by phone.
Anderson views maintaining the accuracy of the library database as a top priority in the conversion to the new system. "Library catalog information is at the heart of MIT because it represents and provides access to much of the intellectual capital of the Institute," he said.
Culminating the changeover will be the development of a client/server system for the Libraries. That part of the new system, also a joint effort between MIT and the Newtonville, Mass.-based GEAC, will be in place by next summer, Anderson said.
The current system is host-based, which means that one central computer does all the processing of data for every user of the system, and individual users connect to this machine. Professor of Electrical Engineering and Computer Science Jerome H. Saltzer '61, head of the Library 2000 Group at the Laboratory for Computer Science, described this as the "milking machine attachment mechanism."
By contrast, in a client/server system like the one on which Athena is based, client software on the computers of individual users access data and information from a small number of server machines, and the clients themselves do the actual processing. This more evenly distributes the burden of processing throughout the network, and makes the system more flexible and adaptable to future changes, Anderson said.
The client/server model is an essential part in the planned development of clients for platforms such as the World-Wide Web, Anderson said.