|CORNELL COMPUTER SCIENCE VISION FOR THE NEXT DECADE|
In the great American universities of the 21st century it must be possible for any student to bring to bear on any subject the ideas and technology of computer science.
1. The University in the Information Age
The information revolution is transforming universities, for it goes to the heart of what universities are about: the creation and dissemination of knowledge. Wise investment by academic administrators will enable a few universities to lead the country into this information-based future. Thus, the information revolution will not only fundamentally transform universities, but it will reorder them in the minds of the nation's leaders. The premier universities will be energized and will prosper by their influence and impact.
Rapid intellectual advances over the past fifty years have created a new field of intellectual endeavor, called computer science in North America and informatics in Europe. The field not only encompasses a large body of knowledge about algorithms, processes, information, communication, languages, and systems but also concerns a new paradigm for representing and acquiring knowledge. In this paradigm, computer programs embody theories. For example, a program might embody a theory of speech recognition, visual processing, or economic behavior theories that are important in psychology and economics. And, computer programs can embody dynamic models of physical phenomena, thereby serving as a basis for a computational science that augments experimentation with a new means of learning about physical reality. These computational simulations have already led to remarkable results, such as the discovery of the exact difference in the speed of rotation between the molten core and the earths mantle.
Unlike other academic disciplines, progress in computing seems to be driven by industry start-up companies and huge corporations, with Microsoft and Intel being the most conspicuous examples. Engineering developments in hardware processors, memory chips, and disk drives and software technology appear so rapidly that it might seem as though universities can not contribute to the field and should not even be trying.
But the role of universities has been transformed not diminished by industrial interest in computing. Look carefully and one finds that the technical insights and innovations frequently originate in universities; it is only their development and commercialization that requires the scale of resources that corporations can provide. Universities may be less involved in the final steps of a commercialization than before, but they remain critically involved in innovation. Moreover, ideas that do originate in corporations are increasingly developed by "wizards" who are younger and closer to their university experiences than technological leaders of the past. For example, a recent Business Week article explained that Microsoft's strategic reorientation toward Internet-based products in 1995 was triggered by Cornell computer science graduate Steve Sinofskys visit to his alma mater .
The young are playing a driving role in the information revolution: they sense the intellectual ground moving, they want to understand the changes, and they want to take part in the changes. For Cornell to continue to attract and excite the best of these curious young minds, Cornell must be seen as one of the leading centers.
Periods of rapid change are opportunities in which decisive action can have lasting effectsand inaction can cause irreversible losses. As Cornell President Hunter R. Rawlings III said;
Ezra Cornell and Andrew Dickson White ... willed into being a new kind of American university. It has, to a unique degree, fulfilled the destiny of the American research university in the twentieth century; it offers instruction, and pursues research, in virtually every field of inquiry .
The CS department has given this Cornell administration the means to "will into being" the model American research university of the information age.
2. Cornells CS Department Today
Since its founding in 1965, Cornells Computer Science department has been ranked among the top five in the country, together with MIT, Stanford, Berkeley, and CMU. Only three or four departments at Cornell are so highly ranked. University figures show that Cornell CS is also the fourth-highest department in the University in generation of overhead revenue. By all kinds of other measures, from Putnam exam winners and Cornell Merrill Scholars among our undergraduates, to the GRE scores and NSF fellowships of our PhD students, to the awards won by our junior and senior faculty, Computer Science registers as one of Cornells outstanding education and research environments.
The CS faculty is relatively young (half under 40 and only 15 out of 26 tenured), highly collegial, and cohesive though diverse, we have trust and common vision. Two other features distinguish our department sharply from the other top computer science departments. First, we are substantially smaller, about half the size. Second, we are undergoing a change that has few parallels among other top CS departments. For our first 20 years, we emphasized fundamental principles of computer science and were perceived as a "theory department". For the past ten years, we have acted decisively to balance fundamentals and high-impact applications, and theory and applications are now equal partners in Cornell Computer Science. For example, here are the rough proportions of our faculty with primarily applied interests, a decade ago and today university departments do not move much faster than this:
1986-87: 7 out of 23
1996-97: 14 out of 26
With regard to research, we already have ties with many other units at Cornell (see the table below), and we are one of the few departments that belongs to two colleges: the College of Arts & Sciences and the College of Engineering.
The Computer Science Department believes that the science behind computing has become so deep and information technology so pervasive that they are relevant to every subject in the university. This discipline epitomizes the notion of an enabling science. The traffic of students and researchers carrying ideas and techniques from our courses into every discipline, department, and institute is evidence of an academic infrastructure for the broad computer science community that President Rawlings challenges us to create.
3. Beyond 2020A Computer Science Community at Cornell
As the decades unfold, the relevance of the ideas and methods of computer science to almost every subject will lead to computing expertise among the faculty in many Cornell departments. This expertise may be in the form of computer science Ph.D.s who have branched into other areas, or of faculty trained in other disciplines who have had substantial education in the methods and outlook of computer science. Our situation will be similar to that of mathematics, which is a key part of the research of many scientific and engineering fields. But computer science can be expected to have a broader scope than mathematics, influencing, in addition, fields like linguistics, psychology, library science, music, and law.
This movement of computer science to other disciplines has begun: already at Cornell we see substantial computer science expertise in the Johnson School of Management, Mathematics, ORIE, and EE. However, the full-scale influx of computer science into other disciplines may take another 2050 years. Computer science is still young. There are not many PhD-level computer scientists, relatively speaking, and those in academia usually prefer a computer science department as their home. Moreover, the turnover of faculty in other departments is slow, so it may be difficult to add computer scientists even when appropriate candidates are available.
The building of a strong computer science community throughout Cornell will depend on the strength and breadth of the Computer Science Department. A strong, broad department will attract good people in other areas as has already happened in ORIE. A strong, broad department will foster interest in computing in other areas. It will be a catalyst for integrating these other departments indeed, for helping to break down departmental barriers.
We believe that Cornell's stature during the next century will depend in part on how well Cornell as a whole embraces computing and information technology in research, education, and administration. In turn, this will depend on the strength and breadth of the Computer Science Department.
4. Goals for the Next Decade
Our national reputation rests on our outstanding educational programs and research: Bachelor of Science (BS), Bachelor of Arts (AB), Master of Engineering (MEng), and Doctor of Philosophy (Ph.D.). Our Ph.D. program was ranked 4th nationally by the latest National Research Council study. Our MEng program is relatively new, and these programs are as yet unranked. We consider our Bachelor of Science program second only to MIT and expect to have that confirmed by a program review scheduled for Fall 97.
Our research reputation today is derived from our facultys fundamental contributions in theory, computational mathematics, programming languages, information retrieval, and distributed computing. Our reputation is also based on groundbreaking texts in these core areas of CS. We intend to remain preeminent in these. But, as computing and information technology has penetrated deeper into all fields of human endeavor, the field has become broader. Thus, to increase our standing, or even to maintain it, the department must broaden as well.
Goal 1: Increase the quality and breadth of our educational and research programs.
Becoming broader will also increase our impact at Cornell. The vast change of scale predicted for every part of computing and communications from speed of processors to capacity of memories, from penetration of networks to size of software systems, from size of datasets to bandwidth of optical fiberswill lead to new developments in all fields. To take full advantage of these developments often requires computer science expertise; at the same time, the problems of these fields create new and interesting problems for computer scientists to explore.
The synergistic effect of interdisciplinary work for the advancement of computer science cannot be overestimated. For example, the biological sciences, crucial in the future, have large computational problems and huge amounts of data. (MIT has apparently recognized this and has targeted Biology and Computer Science as the two fields it will develop in the coming decade.) In Cornell's medical college, research depends on computation. In Cornells Johnson School of Management, there is growing interest in computational finance; economic theories now depend on sophisticated simulations for experimentation and prediction. In physics, fluid mechanics, dynamical systems, and the like, experiments with larger and larger amounts of data require more and more computer science expertise.
Goal 2: Increase our impact across the university in education and research.
The Cornell community would also benefit from broadening the CS Departments educational mission. Today, we do an outstanding job of offering rigorous technical courses that appeal to technically-minded students throughout Engineering and Arts & Sciences. Our research experience for undergraduates program involves 40-50 students per semester, who participate in departmental research projects. But, increasingly, we are asked by other departments and units to bring the information era to non-technical students. These students seek courses in practical aspects of databases, in cognition, in tools for computational science, and in the societal impacts of computers, for example. We must make the subject of information one of the basic kinds of literacy achieved by every graduate. And this leads to a clear subgoal:
Make Cornell conspicuous nationally for the innovative ways in which it extends fundamental computing ideas not just computers to all its students.
We hope to introduce several new courses in an exploratory manner, based on ongoing discussions with faculty and students. Our record for outstanding teaching means that the courses could be exciting opportunities for hundreds of Cornell undergraduates as well as being vital to their ability to function effectively in the information age.
Our research is also relevant to increasing our impact; this leads to a second clear subgoal:
Extend our research infrastructure to provide new information resources for our courses and eventually, with help from Cornell Information Technology, provide this for many courses at Cornell.
The explosive growth of the Internet and the World Wide Web over the past three years underscores the dramatic need for better information resources. As we move into an information-based society, these needs are accelerating. Easy and timely access to the right information is becoming central to nearly every aspect of daily life, from business and health care to education and entertainment. Yet the information resources provided by the Web merely whet ones appetite, providing only a glimpse of what it might mean to have easy access to information. Our goal is to enable qualitatively more effective information resources, to develop a substrate that can provide access to anything, anytime, anywhere. To achieve this goal, we plan to: (1) do fundamental new research, (2) develop integrated daily use by providing new information resources for our courses. In coupling this research closely to our education program, we hope to improve the quality of our courses and to gain valuable experience for our research by putting our ideas to the test in real systems.
In the past, the CS Department has also helped the University in providing more than academic infrastructure. We have helped identify and acquire state-of-the-art hardware for the University as a whole; we brought the ARPA net to Cornell; we provided early Unix experience, which then spread to other departments. Today, we maintain Sun computers for the entire university. We provide leadership at the Theory Center and in Applied Mathematics. In the future there will likely be more calls upon us. We intend to answer those calls and thereby support President Rawlings goals of enhancing the academic community.
Goal 3: Provide leadership in building the wider computer science community at Cornell.
5. Plan for Reaching the Goals
Meeting these goals of increasing our preeminence nationally and increasing our impact at Cornell will require:
Strengthening the core areas of CS
Cornell's greatest strength in computer science has always been the eminence of its faculty, almost all of whom are nationally recognized leaders in their research fields. Traditionally, the department has been especially strong in core areas such as theoretical computer science, programming languages, information retrieval, algorithms, and scientific computing?the five most senior people in the department are in these areas. Although interdisciplinary research is becoming important, these core areas remain essential, and the department must move aggressively to maintain its leadership in them as gaps are created by retirements and other departures.
Expanding in key applied/interdisciplinary areas
In distributed systems and computational science, we are as strong as any university. But the time is ripe to strengthen and broaden areas in which Cornell and the world need research in order to make the most effective use of computing. Below, we briefly outline four such areas.
1. INFORMATION AND MULTIMEDIA TECHNOLOGY. Information technology is the science of organizing, manipulating, and searching data on the huge scale now available to millions via the World Wide Web. In the past, computing technology emphasized textual data, and this will remain important in the future. More recently, the field has broadened to incorporate multimedia data, notably images, video, and sound, and several of our younger faculty members are involved in these areas. We also to build on out prototype distributed digital library, NCSTRL.
2. COMPUTATIONAL SCIENCE. Building on our strength in computational science, we would like to move decisively into more interdisciplinary areas. For example, algorithmic methods will be necessary in nanofabrication (as our joint work with EE shows), and simulation will be important in materials science. Some of the great discoveries that lie ahead in biology will be made possible by large-scale performing computations of both discrete (e.g., genome sequencing) and continuous (e.g., protein folding) nature. Indeed, besides computing and communications, biology is the other great scientific field in which revolutionary developments in the 21st century can be predicted with certainty.
3. SECURITY AND SURVIVABILITY. Issues of security in networks were most familiar in the past through questions of cryptography, including the public-key cryptosystems originated in the late 1970s and the many developments that followed from them. For two reasons this field is destined to explode in new directions. First, the information and control processes on which society depends are fast becoming globally available, raising frightening security concerns that we are only beginning to understand. Second, just around the corner, financial transactions are destined to move to electronic form. "Network security" and "national security" are merging into one, as the pressing security issues evolve from mainly privacy of communications to the very operation of the economic and industrial system on which we depend. Our department already plays a prominent research and policy role in this area.
4. NETWORKS. Computer-communications networks transport information representing everything from letters and photographs to bank transactions and educational lectures cheaply and nearly instantaneously. A ubiquitous computer-networking infrastructure has the potential to fundamentally transform society. For example, ready access to information over the Internet not only impacts the publishing, journalism, and broadcasting industries but also impacts the nuts and bolts of the democratic process. Similarly, electronic commerce over a ubiquitous infrastructure can touch the life of every citizen. The speed at which networking advances, as well as the direction it takes, will depend on the research being performed at places like Cornell and on the education that future leaders receive in this field. We are in a strong position here because Cornell has wired the campus. We intend to build aggressively on this base.
These four areas are interconnected, so progress in any one can affect the others. Better computational techniques will enable us to deal with the mass of information available, better networks will allow us to move the information around more quickly and seamlessly. Better security will give us greater confidence in using the networks. In addition, progress in any of these areas has the potential for having an immediate impact on society. The phenomenal growth of the World Wide Web is perhaps the most obvious example of just how immediate that impact can be. To take an example from a different domain, computational biology techniques play a key role in mapping the human genome, which in turn is expected to lead to a better understanding of human diseases.
Our ambition is not to create four new units within the department but to make the applied side of our department a cohesive and powerful group that is second to none in its influence in bringing academic computer science to bear on problems faced by science and society. Our strength in these areas will help many researchers across the university and will keep Cornell current in the fast moving computing field.
Ten years from now, we intend to still be in the top 5 nationally, despite steadily intensifying competition, and also to maintain the closeness and spirit of cooperation that has made this a pleasant environment in which to work. Finally, we intend to be more influential and visible within Cornell and more of a draw of students and researchers to Cornell than ever before.