THE COMPUTER REVOLUTION

The College's new Center meets needs of both faculty and students and promises to make the computer part of today's liberal learning

ONE HUNDRED years ago the full effects of the Industrial Revolution, with all the economic and social changes that it brought about, were just beginning to be felt in America. Today America is at about the same point in a comparable upheaval - the Computer Revolution - which promises to have an equal or greater impact.

The danger in any revolution is that the forces that generate it be overestimated or underestimated. The Industrial Revolution did not end poverty and ignorance and create a perfect human society as some claimed it would. But it did increase individual productivity, it did raise standards of living, and it did lessen the sheer physical drudgery of many tasks.

Today high-speed computers are revolutionizing science, industrial production and planning, military strategy, transportation, and record-keeping of all kinds in much the same way.

Problems that humans would require days to figure out are solved in seconds; computing tasks that seemed so insuperable that they were abandoned are easily programed and solved.

Again some people are forecasting new computer-created Utopias. Others feel that the machines will stifle creativity and so dehumanize existence as to make it unbearable.

Neither position is valid. The computers are tools. They are marvelous tools, to be sure, but they are still only capable of tasks that men design for them and they are incapable of the purely speculative kind of thought that leads to the creativity that sculpts a "David." Or designs a computer.

Obviously, if the computers are to occupy this important place in the future, some meaningful contact with them must be a vital part of any liberal education, if only, as one mathematician put it, "to protect ourselves from them."

His remark answered those who acknowledged the need for such instruments in the Thayer, Tuck and Medical Schools and perhaps even some science departments, but wondered how such technology could be a true part of Dartmouth's commitment to the liberal arts.

The point was that the daily life of everyone is being drastically influenced by computer concepts and any such pervasive force demands a place in a curriculum that hopes to teach about man and the world.

Some years ago, stimulated by the Computer Revolution, Dartmouth began a careful study of needs and opportunities. About 75 per cent of all Dartmouth students take at least two math courses and about 90 per cent take at least one. There was both a need and opportunity in instruction, then. In addition, faculty and student research was increasing tremendously and computers were opening up new approaches to many problems. Here too were both needs and opportunities.

In 1959 an LGP-30 was obtained for experimentation, teaching, and research. Later the Tuck School got an IBM 1620 and the Thayer School, while using both these instruments, also experimented with smaller, individual computers.

From these experiments, from studies of the needs and surveys of the equipment available, a group headed by Prof. John G. Kemeny, Prof. Thomas E. Kurtz, and Dean Leonard M. Rieser evolved a plan.

In effect, it allowed Dartmouth to "leapfrog" in the computer world. The previous installation was quite ordinary. Similar systems could be found at many institutions and many had vastly superior installations. The new computation center puts Dartmouth in the forefront among educational institutions.

Professors Kemeny and Kurtz feel that its users may well be getting the fastest service on research that any institution can offer. They are also confident that few - if any - can offer such quick and easy access to a high-speed computer. "It's almost as if each faculty member and student had a high-speed computer of his own."

In their planning they had a new breakthrough in computer technology to work with - "time-sharing." In addition, the National Science Foundation made a $300,000 grant for equipment and the College received special cooperation from the General Electric Company in making purchases.

Peter Kiewit '22, president and director of Peter Kiewit Sons' Company, a world-wide heavy construction firm, is giving $500,000 toward the purchase of equipment and construction of a building to house it. The building, to be located near the Albert Bradley Mathematics Center, will be known as the Kiewit Computation Center. The center is temporarily housed in College Hall.

Time-sharing permits widespread use of computers because it eliminates the major bottleneck in computer usage: getting the problem in and the answers out. The computers require only seconds to solve relatively complicated problems, but before time-sharing a user could easily spend a week getting a solution even after he had his problem's program worked out. For instance, someone with a problem would take a set of punched cards to the computation center on campus and present them to an attendant. The program - let's say it is a relatively simple one involving only a few thousand computations - is grouped with others in batches. Eventually they are fed into the computer, processed and answered. The user then returns to the center - perhaps the following day — to get the finished computations.

Perhaps some minor error in the program became apparent. This would have to be isolated, corrected in a process called "debugging," and the whole process repeated. Few programs are entirely free of bugs the first time around.

This wasted a great deal of the user's time. It also wasted the computer's time.

Obviously, widespread use by students, faculty, and administrative offices could not be obtained this way with a single computer. And getting enough computers to handle all the potential business would be prohibitively expensive.

The study group examined the available equipment and — in collaboration with General Electric's representatives - devised a system to meet the College's needs. All the equipment consisted of "off-the-shelf" items. None had to be customized at considerably greater expense.

Its basic elements were two computers, the GE 235, a high-speed computer capable of handling virtually all the problems likely to arise at Dartmouth, and a GE Datanet 30, a new machine described as a "communications-oriented computer."

Another key element was the GE Mass Random Access File Memory, which can store up to 18 million characters of information as magnetic spots on 16 magnesium discs.

To make the computer system as easily available as possible, the teletypewriters that feed it problems, the "inputoutput" stations, were placed strategically about the campus. The stations were installed in the Thayer School, the Tuck School, the Medical School, the various undergraduate science and social science departments, at the Center itself, and in the administrative offices. Even Hanover High School has a station. Sixteen were planned in the initial phases, but it soon became apparent that more would be needed. Twenty-two are now in operation and provision has been made for more.

But how does it work? Let's take two simple, easily understood hypothetical cases. Professor X of the Economics Department is studying New Hampshire banks and has compiled considerable data on them which he has previously fed to the computer and which is stored in the memory disc. Student Y is studying with Professor X and using his data. Professor X wants to know the average rate of "real interest" that each of the banks charges; Student Y wants to know the total capital assets of the banks.

Let's say they pose their questions simultaneously from separate input-output stations. Professor X gives the computer appropriate instructions. The questions are fed into the Datanet 30 which monitors the 235's work. The monitor orders the memory disc to summon up the data already fed to it. When ready it is fed to the computer with the questions and the order given to proceed with the computation. Student Y's problem was stored momentarily in the Datanet 30, then transferred to the memory unit. Professor X's problem is considered by the computer for a few seconds and computations made, but perhaps it requires more than a few seconds for complete computations. The Datanet 30 then orders the computer to run Student Y's problem which consists of merely adding a long list of large numbers. This is solved quickly, the answers returned to the memory unit, and the computer returns to consideration of Professor X's problem. The Datanet 30 then summons Student Y's answer and orders it printed on the teletypewriter he is using. Seconds later, its computations done, the computer returns the answers to Professor X's problem to the memory unit and they go back via the Datanet 30 to his teletypewriter.

Meanwhile perhaps several other stations are being serviced simultaneously since the computer considers and works on each problem in turn.

The easy, quick access provided by time-sharing has another major advantage. The "debugging" process poses many difficulties and slows the computation considerably. Previously when an error in data or instructions was discovered at the end of the computation, the user had to return to his office or room, find the error, correct it, and resubmit the problem.

To illustrate "debugging," let's return to Professor X and his relatively simple problem. In the long list of banks whose "real interest" charges he wishes to know, he finds one that is charging 76 per cent interest. He is sure this is in error and re-examines the data. There he finds an error in the income from interest for that bank. He then types new instructions to the machine which involve only the line on which the error occurs. The Datanet 30 then tells the memory unit to erase the erroneous line and substitute the correct data.

The program is then recomputed and the correct answer given.

There are many analogies to timesharing. It resembles in many ways the technique a chess player uses in playing several games at once. He doesn't play one game all the way through and then start on another. Instead, he moves one man at a time on each of the boards and keeps all the games going at once.

The Datanet 30 is the monitor which listens to all the separate requests for information, keeps track of who has called, and allocates the computer's time efficiently.

With the computer network at hand and operating, a second phase - the education of its users - began.

The Mathematics Department felt that a separate course on computers was neither practical or desirable at this level. Instead, they chose to include several hours of instruction in the basic mathematics courses. Expert individual instruction is also available at the center if needed.

In addition, Professor Kemeny was given an $84,150 grant by the National Science Foundation for the first year of a two-year program involving experimental undergraduate instruction in computing.

One of his first steps was devising a simple language through which man and machine could converse. He called it BASIC (for Beginners All-purpose Symbolic Instruction Code). This combined simple, easily understood English words and mathematical symbols. Each of the words and symbols has a more precise meaning than it would in conversation because, as the BASIC manual says, English is rich "in ambiguities and redundancies, those qualities which make poetry possible, but computing impossible." BASIC is simpler than some of the other computer languages such as ALGOL or FORTRAN, but the computers will also accept programs in these latter two languages.

For faculty members and others a short instruction period was conducted last summer just after the equipment was installed. A second course for faculty and researchers is scheduled to begin in early November.

One veteran of that first faculty course was Prof. Hugh S. Morrison '26, an art and architecture historian who describes himself as probably the man on campus "least likely to benefit from a computer."

To test the computer, he posed a calendar problem: Give the day of the week of the Massacre of St. Bartholomew's Day in Paris, August 24, 1572.

Professor Morrison said: "The darned thing talked back to me in three seconds, reporting that allowing for the change in the Gregorian calendar in 1752 to the reformed or 'New Style' calendar, the massacre occurred on a Wednesday."

Afterward, in reporting on the instruction, he said: "The course profoundly affected the thinking of all of us. This is the most important thing - much more important than the machine itself. Of course, we know that it is the brains behind the machine that make these miracles possible. Nonetheless, it is a weapon of such power that all intelligent men and women everywhere should know the kind of things it can do. Once we know that, we can devise ways to make use of it. ...

"This machine of ours (emphasis his) is a friend."

Prof. Thomas E. Kurtz (standing) is director of thenew computation center, temporarily in College Hall.

Students using one of a half-dozen input-output stations centrally located in theformer Commons dining hall. This cluster is part of a network of remote stationsaround campus now numbering 20 and scheduled to be enlarged in the near future.

Sidney Lees (top), Professor of Engineering, and Brian Walsh '65, an engineeringscience student, using remote stations ina striking example of the double valueof computers to faculty and students.

William R. Zani '64T, Supervisor of the Computation Center, shown with the Data-net-30 which monitors all incoming calls and allocates use of the computer's time.