Model 3: An Independent Unit (Division) for CS&E

In addition to creating a CS&E Graduate Group, this model would lead to the creation of a new, independent organizational structure for CS&E. Model 3 for CS&E at UC Davis would be the establishment of an independent structure responsible for coordinating all educational efforts in CS&E. Potential Problems with a Unit for CS&E, Coupled with an Undergraduate MAJOR in CS&E. Concerning the profile of a CS&E faculty member, close ties with at least one application area and a certain degree of sophistication in computer science, preferably coupled with, for example, a background in applied mathematics or statistics, are essential. Concerning the student body, it is reasonable to assume that the number of students sufficiently prepared to master demanding material concerning the development and study of CS&E methods is relatively small. The amount of training enabling a student to take CS&E courses--the kinds of courses that would not merely teach the application of tools but actually deal with the underlying concepts and their study--is substantial. This is particularly clear for material concerning scientific computing (including, as its core, numerical analysis and numerical linear algebra), multidimensional data mining algorithms, or optimization methods. Thus, one should be concerned about the number of undergraduate students who would qualify for and benefit from CS&E courses. Hence, the establishment of an undergraduate major might become a long-term goal of this Initiative, but there would be serious problems one would have to overcome: (i) defining a pool of CS&E courses large enough to provide the number of credit hours needed for a major; (ii) creating mechanisms that would link the CS&E course material to applications in which the CS&E methods taught are relevant; and (iii) ensuring that the courses of such a CS&E major would not be de facto duplicates of existing courses. Students with an already established background in scientific and engineering applications of CS&E, computer science, and possibly mathematical methods would be primary target groups for CS&E course material. To equip undergraduate students with the required high degree of sophistication prior to their taking CS&E courses, one would have to revise existing undergraduate curricula and requirements substantially to prepare students accordingly. An undergraduate CS&E major without a serious connection to specific applications of CS&E might duplicate, to a large extent, course material from Computer Science, Mathematics, or Statistics. This is not desirable. CS&E certainly should not lead to a ``watered-down version'' of these three existing programs, or others. Thus, the inherent interdisciplinary nature of CS&E is an obstacle for the establishment of an undergraduate major. A CS&E curriculum, by its very nature, should expect some knowledge of application-specific problems and basic requirements concerning possible computational solutions. Given the wide range of application domains, what would be common to the CS&E curriculum, faculty, and students in a CS&E undergraduate major? Most likely, the commonality would be rather generic mathematical and computer science methods and problem solving skills, applicable to a variety of computational problems encountered in science and engineering disciplines. On the other side, there is the danger that a major might focus on highly similar applications, and one would have to ensure that such a major would not just teach software packages and tools or evolve into a ``second Computer Science Department.'' Unless the faculty of an independent unit responsible for CS&E course delivery are focused on highly similar aspects of CS&E, there would hardly exist any intellectual core to hold this unit's faculty members together. The commitment, then, should be the use and development of computation as a tool to enable scientific discovery through computation. Potential Advantages of a Unit for CS&E, Dedicated to an Undergraduate MINOR and Graduate Education. An alternative to the establishment of an undergraduate major in CS&E within a department-like structure would be the development of CS&E to support double majors, minors, designated areas of emphasis, etc. This could, possibly, be implemented within existing departments representing application areas of CS&E or within Computer Science. An administrative structure facilitating double majors, minors, designated areas of emphasis, etc. would be needed, and the campus would have to implement mechanisms that would encourage undergraduate students to pursue such dual proficiencies. Improved future job competitiveness would be an important argument to attract student attention to such programs. Students would major most likely in a science or engineering field, computer science, or mathematics, with a specialization in CS&E. A divisional administrative structure to support undergraduate minors and a graduate major in CS&E should be considered strongly. For example, this model could be implemented by the creation of an independent Division of CS&E hosted in Computer Science or--and this is the option favored by the CS&E Committee--in the College of Engineering. This is analogous to the Division of Statistics, hosted in the College of Letters and Science. This model ensures that the educational and intellectual interests of other colleges are protected. Courses, minors, and majors that such a Division would offer should be accessible to the general campus student body as much as possible. This aspect should be paramount concerning course delivery--regardless of what type of organizational structure for CS&E is chosen in the end. If universal access to CS&E courses were guaranteed, such a division would be attractive to the campus at large. If a Division of CS&E were eventually placed in a particular college, e.g., Engineering, it would be pivotal to ensure that students from across campus would have equal access to CS&E courses. Major resources would have to be committed to a new Division of CS&E, in terms of space, computing resources for educational laboratories, and administrative and technical support. The College of Engineering and the campus would have to make significant investments to make this model work. Placing a Division of CS&E in the College of Engineering would locate it in the same college as Computer Science, which should be positive in the context of course coordination and minimizing the danger of course material overlap or duplication. The creation of a Division within an existing department, i.e., Computer Science, could potentially lead to internal conflicts as a result of competition for resources. One would minimize this danger by creating an independent Division of CS&E in the College of Engineering. Some Committee members see a danger for possible duplication of efforts. The mission of Computer Science might, in the end, have serious overlap with the goals set forth for the CS&E Initiative. This should be avoided. Furthermore, one has to assume that hiring priorities would have to be changed to some degree by departments heavily impacted by CS&E expansion. To foster more widespread and more enthusiastic support concerning expansion in CS&E areas, one might promote joint appointments: Model 2 could foster joint faculty appointments and could thereby make the hiring of CS&E faculty attractive--both to the department(s) where CS&E appointments are concentrated and to the application domains in which CS&E faculty would be appointed jointly. The CS&E faculty members would define and coordinate a graduate program and, most likely, minors or emphasis areas in CS&E together with faculty from all disciplines with an interest in computation.