25th Anniversary Talk

25th Anniversary Celebration of Columbia’s Computer Science Department

DEPARTMENTAL RESTROSPECTIVE

JOSEPH F. TRAUB

Before the establishment of the Computer Science Department in1979, there were computer science efforts in the Mathematical Statistics Department in Arts and Sciences, and in the Electrical Engineering Department in the School of Engineering.  In the academic year 1978-79 Peter Likins, the Dean of the School of Engineering (now President of the University of Arizona) persuaded the University to eliminate both these efforts and to create a Computer Science Department.  The creation of the new Department was strongly supported by Columbia's central administration.

At the time I headed the Computer Science Department at Carnegie Mellon University, universally ranked together with Stanford and MIT as one of the premier departments in the world.  Peter approached me to be the founding chair of the new department at Columbia.  I agreed to come under certain conditions which included:

  • a new building for the Department
  • faculty positions -  Likins started the Department with thirteen slots and said "don't ask for any more".
  • a teaching load competitive with leading computer science departments at  private universities

Likins agreed to all requests and I came to Columbia University on July 1, 1979. The new Department had four tenured professors (Theodore Baskow, Jonathan Gross, Stephen Unger and myself).  The decision was made to recruit the best possible new PhDs , as well as a few outstanding senior faculty, which required that the existing non-tenured faculty had to be asked to leave.

The following people, who are now all senior faculty joined the Department in the early years:

    Peter Allen
    Steven Feiner
    Zvi Galil
    Gail Kaiser
    John Kender
    Kathy McKeown
    Sal Stolfo
    Henryk Wozniakowski
    Yechiam Yemini

It was decided to import some of the Carnegie-Mellon principles which included:

  • a strong commitment to research and teaching quality
  • faculty collegiality
  • treating students as colleagues
  • a Black Friday meeting every semester to evaluate all PhD students

Many people have asked me why this evaluation meeting is called  Black Friday, especially when at Columbia its usually been held on Thursday.  The reason is shrouded in the fog of computer history but I'd like to share it with you.We have to return to CMU in 1971 where I met weekly with a faculty-student committee which was revising the PhD program.  One Saturday night my wife, Pamela, and I were watching Chilly Billy's Thriller Theatre on TV and we saw a Grade B horror movie about the dead returning every hundred years called Black Sunday.  The following week the Committee decided to create a twice-a-year meeting to monitor the progress of our PhD students and scheduled the first meeting on a Friday.  When I notified the CS Department, I casually referred to this as the Black Friday meeting.  The name stuck and when I brought the idea to Columbia, we continued to use it.  

Much had to be accomplished in the creation of the new department.  Peter Likins was marvelously supportive at all times.  Here are some of the things we  had to do.

The faculty had to be built almost from scratch.  One of the characteristics of computer science, as opposed to many other disciplines, is that there are many more available positions than absolutely top people to fill them.  It was true at Carnegie in 1971; it's true today.  That is healthy for the discipline but makes it difficult for faculty recruiting.  For every appointment, the new Department had to compete with top academic departments and top research labs.  Nonetheless, superb new faculty were hired.

At the same time there was huge student demand for courses.  Two thousand students took courses in the first year, 2600 the next year, and 4200 a few years later.  There were as many as 200 students per course.   We were trying to hire the top young PhDs and yet they had to teach classes of  200 students.  The Department started to hire lecturers to help with teaching , especially for the beginning courses.   We hired lecturers who loved to teach and who interacted well with undergraduates.

The Department started Bachelor's, Master's and PhD programs.  It taught computer science to all of Columbia University so there were majors from the Engineering College, Columbia College, Barnard, General Studies and the Graduate School.  Jonathan Gross served as Vice-Chair.  Among his responsibilities were creation of the curricula and oversight of the various Bachelor's programs.

Obtaining funds was crucial to building Departmental research.  In the first year, IBM gave the Department a six hundred thousand dollar gift.  Pivotal to giving the Department a big push was a very large contract from DARPA, the Defense Advanced Research Projects Agency.   DARPA monies had been critical to the building of the "big three":  Carnegie-Mellon, MIT and Stanford.  Bob Kahn, then the Director of DARPA's Information Processing Techniques Office, decided to give major funding to two more Departments, Berkeley and Columbia. The DARPA funding was crucial to the Department's early years. In addition to providing funds for research staff, research equipment, and PhD students, it provided funds for equipment and staffing of a computer facility.  Furthermore, it gave the Department a vital connection to the Arpanet.

New York City was a major plus in recruiting faculty and students but could sometimes make matters more complicated.  The Arpanet connection provides a good example.   In 1979 Columbia did not have an Arpanet connection. Parenthetically, although its hard to believe today, the entire Engineering school's only computer was a PDP11/45.  How to connect to the Arpanet?  Connecting via NYU would have been the obvious solution but the density of wires and pipes in midtown made that unfeasible and we had to come up with a more complicated solution.

As I mentioned earlier, there was a University commitment to give the Department a new building.  The boutique firm of Kliment and Halsband was chosen as architects.  The building won the AIA Honors Award, which, for an architect, is the equivalent of an Oscar or a Pulitzer Prize.  To celebrate the new building, a convocation was held in 1983.  There were talks by distinguished leaders in computing and Herbert Simon received an honorary doctorate.

In 1979 Columbia University was fairly late in starting a Computer Science Department.  The Department felt it was important to demonstrate to the University the centrality of the discipline.  One of the ways this was accomplished was through the “Columbia University Lectures in Computer Science".  The distinguished and diverse speakers included sociologist Sherry Turkle, computer animation pioneer John Whitney, and Douglas Hofstadter, author of the Pulitzer Prize winning book, "Goedel, Escher, Bach".  The lectures always drew large audiences.

In 1984 the Department was five years old.  It had twenty four faculty positions (Peter Likins' limit of thirteen had long been passed).  The PhD program had 60 students and there were 6 to 7  million dollars a year in outside support. A state-wide competition led to Columbia's designation as the New York State Center for Computers and Information Systems.  IBM gave a generous gift of  4 million dollars in money and equipment.  The first stage in the building of  the Computer Science Department had been completed.