History of the Department of Computer Sciences
at
Purdue University

John R. Rice and Saul Rosen

This account is a condensation and updating of two articles by John R. Rice and Saul Rosen in the book Studies in Computer Science: In honor of Samuel D. Conte (R. DeMillo and J. Rice, eds.) Plenum Pub., 1994. The articles are The Origins of Computing at Purdue University , pages, 31-44 and History of the Computer Sciences Department of Purdue University , pages 45-72.

1. INTRODUCTION

The first Department of Computer Sciences in the United States was established at Purdue University in October 1962. There are three natural phases in its history. In the 1960s the effort was to define courses, degree programs, and indirectly the field itself. The 1970s saw the department's maturation and growth into a typical university department. The 1980s started with a series of crises, some nationwide and some internal to Purdue, which eventually gave the department a considerably different character than it had in the 1970s. Table 1 presents a chronology of the principal events and milestones for 1962--1996.

2. THE 1960s: ESTABLISHING THE CURRICULUM

The first task of Samuel Conte as new department head was to hire some faculty and define a graduate program. The course offerings planned were not large -- enough graduate courses for the M.S. and Ph.D. degrees and an undergraduate service course in programming. In the very first year, there were five teaching faculty, Samuel Conte, Richard Kenyon, L. Duane Pyle, Robert Korfhage, and Saul Rosen. Although not all of the faculty taught full time, they could offer over 20 courses a year, which was ample to support the planned program. In 1963 there were three new faculty members: Richard Buchi, Walter Gautschi, and John Steele. The following year John Rice was hired.

The new Mathematical Sciences building was completed in 1967, and the department (along with statistics) moved there from the Engineering Administration building. The Computer Sciences Center occupied the two floors below ground. The department occupied the fourth floor, which was substantially larger than the previous space and also much nicer.

The Department of Computer Sciences was a part of the Division of Mathematical Sciences along with the Departments of Mathematics and Statistics. Initially Felix Haas was head of the division and also head of the mathematics department; the three departments were only partially independent within the division: They set degree requirements separately, but there was only one graduate committee and one Ph.D. qualifying exam system. This arrangement was appropriate in view of the small sizes of the computer science and statistics departments.

2.1. Undergraduate Program

The undergraduate program evolved initially from very sparse courses offerings in programming to a computer science option in the mathematics department to a separate B.S. degree approved in 1967. Conte was an active member of the Association for Computing Machinery committee that recommended the model B.S. degree program known as Curriculum `68. It very closely resembled the degree program at Purdue, which was one of the test beds for developing Curriculum `68. It was not until well into the 1980s that the undergraduate computer science program included the variety of offerings that was common in the other sciences.

3. THE 1970s: MATURATION

By the beginning of the 1970s, the department had completed its pioneering years: Degree programs were established; there was a faculty of 15 and dozens of computer science departments at other universities; and Purdue's department was fully independent. The 1970s were to be a decade of consolidation and maturation. But there were still serious challenges; perhaps the most difficult was hiring faculty. Throughout the 1970s almost every computer science department had unfilled positions for computer science Ph.D.s, as did many major industries.

The faculty shortage was compounded by another trend that became widespread in the 1970s: the change from a mathematics-like discipline (using only paper and punched cards) to a science-like discipline with a significant experimental science component. Some computer science departments originated in engineering and had the experimental component from the beginning, but by the end of the 1970s, most departments, including Purdue's (which originated in mathematics), had started to establish a significant experimental component.

The second serious challenge for computer science departments everywhere in the 1970s was to establish their scientific respectability. Many science and engineering faculties knew about computing only through contact with Fortran programming, and they assumed that was all there was to computer science. Even though the Purdue Department of Computer Sciences was consistently rated in the top ten, it had to reaffirm its permanence and value continually to other parts of the university.

The third serious challenge was the evolution of courses. In spite of repeated reorganizations of courses and the expansion of offerings, it seemed there was always some course that required complete restructuring. The department simply did not have enough faculty to keep all the courses up-to-date at all times, and this situation persists today.

4. THE 1980s: DECADE OF CRISES

The growth and maturation of the 1970s held the seeds for the crises that hit in the first half of the 1980s. These crises stemmed from the numerous major needs and the lack of resources to meet them.

4.1. Crisis 1: Student Enrollment Explosion

The number of entering freshmen majoring in computer science during the 1970s did grow some: About 80-100 entered from 1970-1974. Enrollment then increased to 150 a year from 1975-1977. In 1978 and 1979 the number increased to 200, then 300, and the crisis was on us (see Figure 1). This growth was nationwide. For one year in the early 1980s, a survey showed that 9% of the high school graduates wanted to study computer science. If this percentage had continued, computer science would have had as many students as all of engineering! By the fall of 1981 there were over 500 freshmen starting out in computer science. Earlier groups of students were advancing through the curriculum, and the undergraduate courses overflowed, new sections were added and then they overflowed again.

The administration did offer to increase the number of positions in the department, but that was completely safe. The department already had unfilled positions and having more of them would not increase the number of faculty. The explosion was handled primarily by increasing class sizes. Examples of the extreme situation during that period were: (1) The senior level course in numerical analysis was taught in a single lecture section with about 150 students and a half-time teaching assistant grader, (2) the first-year graduate course in compilers had over 80 students and no grading assistant, (3) teaching assistants had 160--180 students in lower division courses. There was a corresponding lack of computing facilities to support the courses.

4.2. Crisis 2: The Space Crunch

When computer sciences started changing to an experimental, laboratory-oriented discipline in the 1980s, space was needed for departmental computing, teaching labs, and research labs. The result was tighter and tighter packing, and most faculty simply could not engage in laboratory work. For example, in 1983 a faculty member asked the department head for a secretary to help support his work. Heads usually reply that they don't have any money for that, but in this case the department head said, "I understand, let's do it. Tell me where you want the secretary to be and I'll hire one." This offer was safe because the head knew that there was no room in the department to put even one more desk.

4.3. Crisis 3: Establishing Departmental Computing Facilities

The department acquired its first general purpose computer, a VAX 11/780 in 1978. It was the first VAX to be running VAX UNIX outside the developer's sites (Berkeley and AT&T Bell labs). The motivation was the need to have an interactive, time-shared computing environment. It was not practical for the computing center to provide this service on a widespread basis, and it was unwilling to do so for just one department. It was however, inevitable that the department would set up its own facilities as its needs became too specialized and too diverse to be satisfied by a centralized service center. This move was part of the nationwide trend of computer science to become more experimental and laboratory oriented.

This crisis was very real, but it was handled much more smoothly than the others because of the administration's willingness to support this growth. The extent of the changes required is illustrated by Table 2 which gives the operations budget, facilities staff, and installed equipment. In a 10-year period a major new operation was established within the department. This further contributed to the space shortage in the department.

In response, a building was selected in January 1984 for renovation to house the computer science department. The renovation was completed quickly, and the building was occupied in the fall of 1985. The space was of excellent quality and, for a few years, the department enjoyed ample space. However, the need for labs, supporting staff, and research assistants grew rapidly, and by 1989 the packing process was being repeated.

4.4. Recovery

Recovery from these crises occurred in 1985. Moving into the newly renovated Computer Sciences building dramatically improved morale. Not only was the environment greatly improved, but there was an opportunity to start teaching and research laboratories of all kinds. The computing facilities had the type of space (if not yet all the equipment) needed to provide first-class facilities. This physical improvement was accompanied by initial solid evidence that the flood of students was receding. Only a little over 300 of the 1985 freshmen class declared computer science as their major.

5. HISTORICAL DATA

Figure 2a shows the number of A.S. and B.S. degrees awarded and Figure 2b shows the number of M.S. and Ph.D. degrees awarded in computer sciences from 1964-1995. Table 3 lists all the Ph.D. graduates, with their advisers, from 1966-1995. Table 4 lists full-time faculty from 1962-1995 and indicates when they were in the department. Full-time refers to the appointment at Purdue, although several of these faculty members were only part-time in computer sciences. Table 5 lists all the professional staff since the first, William Gorman, was appointed in 1975.

Growth in research in general, and experimental research in particular, is perhaps best illustrated by the increase in research fuding from $447 thousand in 1980 to $3.6 millioin in 1989. Not surprisingly there was also a substantial increase in the number of Ph.D. students during this period and some decrease in the number of M.S. students.

The new space acquired on moving into the Computer Sciences building also allowed the department to establish teaching laboratories. In the first year, 1985, there were four -- two for CS 110 (an elementary personal-computer-based service course), one for CS 180, the first course for computer sciences majors, and one for graduate courses in operating systems and networking. By 1989 this number had doubled.