The Tech - Online EditionMIT's oldest and largest
newspaper & the first
newspaper published
on the web
Boston Weather: 73.0°F | Partly Cloudy

Group suggests changes in academic computing

By Andrea Lamberti

A discussion paper on the future of academic computation at MIT was released last week by the Committee on Academic Computing for the 1990s and Beyond. The draft paper was released in order to foster discussion for a final report which will be presented to Provost John M. Deutch '61 in late May.

The draft includes many possible recommended courses of action for improving academic computing at the Institute. Input from the MIT community in the coming weeks will further narrow the field of possibilities, and will be a major factor in the final report.

The purpose of the committee is to assess the past role of Project Athena and other academic computing at MIT, and to make recommendations for the future.

Committee members collected data for the study by reviewing materials on Project Athena and other computers, interviewing various individuals, surveying students and faculty, consulting with the schools and the MIT Libraries, and appraising academic computing at several other universities, the paper stated. They finished collecting data in late January.

"The committee's view is that this is an accurate capturing of where the committee is in its thinking now, said Gregory A. Jackson '70, director of the study and special assistant to Dean for Undergraduate Education Margaret L. A. MacVicar '65, who chairs the committee.

The educational focus of academic computing

Possible recommendations the committee will make to the provost later this spring "fall along two complementary dimensions," according to the paper. These are "extending and deepening the educational focus for MIT academic computation, and . . . expanding the role of academic computation in the MIT intellectual community through shared resources and user support."

In order to achieve the first aspect of the recommendations, the committee made possible recommendations on the organizational structure of academic computing within the MIT community. Each school would "develop computer-based educational strategies" to support the educational goals of the school and allow students to reach "an appropriate level of professional computing competency upon graduation."

The educational focus of academic computation would also be expanded at the Institute-wide level. The paper gives the example of the "relevant interdepartmental organizations" promoting "the basic computational fluency of all students." The committee also placed emphasis on computing in the freshman year.

Institute efforts for computing beyond the freshman year could be incorporated into General Institute Requirements, the paper stated, such as required HASS-D subjects, Science Distribution subjects, or subjects that cross disciplinary lines such as statistical analysis, mathematical simulation, or access to databases.

"Academic computation at MIT will need to incorporate changes as they occur, and our recommendations . . . aim to facilitate this," the paper states.

The committee found six key areas to focus on for academic computing. One of these, knowledge databases, would exist on large network systems, and could include "curriculum materials for courses, including references to background material" and communication among faculty, students and TA's; "digital libraries with subject indexes and databases that could include mathematical theorems, gene sequences and historical events; on-line support; . . . and on-line magazines," the paper stated.

Another of these areas, intelligent agents, could " `work' for users in a myriad of ways," such as search databases that could answer some questions automatically and [route] other questions to different people; find messages, bulletin board items, and news stories for users; sort mail <>

and other information; manage projects; and keep calendars and perform some scheduling tasks.

Communication and

common resources

According to the paper, many people still use computers today "quite primitively," without extensive communication with other users. One of the visions of Project Athena was to create "a richly interconnected network of individual workstations and powerful servers for storing, printing, and processing information," the paper said.

In the early years of Project Athena, the system did not realize this vision or succeed in advancing learning in a "collaborative sense. . . . Only recently have interaction and collaboration become important objectives in their own right," according to the paper.

Developments in communications networks have decreased the costs, both physical and financial, of data communication, and community members can now "transfer documents and data files among their computers," the paper states.

These developments have contributed to an evolution toward "pluralism" in academic computing, the paper continues. This evolution will eventually require the coherent interaction of hardware, operating system, and applications software.

Coherent interaction is desirable for a community of users, because it facilitates communication among users in a variety of environments. According to the discussion paper, an environment consists of hardware, the connections to networks or shared peripherals, the operating system and application programs.

The paper cites one shortcoming of Project Athena as its "inability to support most commercial software." Project Athena uses the Unix operating system, which does not support many well-known commercial applications.

"The desire to access commercially available programs and the desire to intercommunicate suggest a new level of coherence built around a minimal set of desirable common services -- Community Computation Resources -- that advance our academic computation objectives. Individuals would use commercially available software in the hardware and operating system of their choice," the paper states.

The "essential point to [the] recommendation is not that Community Computational Resources should be available, but that they should be available within individuals' preferred operating systems, within reason. Coherence thus need not imply a single hardware or operating-system platform for centrally supported academic computation at MIT."

In another possible recommendation, the committee suggested a standing Institute policy committee to review academic computing at two-year intervals. It also suggested that over the next five years, schools and departments within MIT develop and implement their own plans for academic computation.

Five implementation

packages suggested

In a discussion of cost and implementation of these possible recommendations, the committee offered five different cost options.

"The first three packages implement our Community Computational Resources recommendations, with varying levels of equipment, user support, and ancillary expenditures." The total costs of these packages range from $15.5 million to $22 million in annual expenditures.

A fourth option would maintain the current level of activity, at a cost of about $12 million per year. The final, pared-down option would cost about $2.7 million annually.

"MIT [must] maintain the current number of workstations, recognizing that in future, as at present, some Departments may use these more than others," the paper states. However, the committee feels that educational applications or user support should not be sacrificed to maintain workstation numbers.