Report of the External Review Committee for the
Computer Science Department
GeorgiaStateUniversity
January 8-9, 2007
Committee Members:
Sartaj K. Sahni, Distinguished Professor and Chair
Department of Computer and Information Science and Engineering
University of Florida
David A. Schmidt, Tointon Distinguished Professor
Department of Computing and Information Sciences
KansasStateUniversity
*Xiaodong Zhang, Robert M. Critchfield Professor and Chair
Department of Computer Science and Engineering
The OhioStateUniversity
*Chair of the review committee.
Executive Summary
We, the three external committee members, visited Georgia State University (GSU)and its Computer Science Department (CSD) on January 8-9, 2007. The self-study document and the instructions we received before ourvisit were well prepared. We met with faculty, undergraduate and graduate student representatives,and key university Administrators, including the Provost and the Dean and Associate Dean of the Arts and SciencesCollege. We were impressed by the wisdom ofcentral administration at GSUto invest in this young department; this investment has already benefited GSU and CSD immensely. We strongly believe that, with continued support of the University and the College, the Department will continue to yield handsome dividends.
The CSD has made impressive progress since it was established in 1999. For example, by the end of 2007, the department will have produced 17 Ph.D. graduates, an outstanding quantity for such a youthful program. Most of the Department’s faculty are actively involved in research activities; those who are not are engaged in supportive teaching and service roles. The faculty is a collegial group, dedicated togenerating national recognition for their Department. The leadership of the recently appointed Chair, Dr. Yi Pan, is appreciated by the upper administration, the faculty and the students.
In this report, we document our observations and evaluation, emphasizingthe Department’s undergraduate and graduate programs, faculty research, junior faculty career development, faculty hiring, staff support, space, and other resource issues.
A. Strengths and Weaknesses
The CSD has several distinctiveassetsthat underlie its accomplishments in the past 7 years. First, in contrast to many other states, the State of Georgia hasinvested heavily in its higher education system. The CDS’s undergraduate program hasbenefited fromGeorgia’s HOPE Scholarships, and the Department’s faculty size and research activities have been strengthened by the Georgia Yamacraw academic program. Second, Atlanta is one of the fastest growing cities in the U.S., providing many cultural, industrial and educationalopportunities, as well as a huge international airport that connects the city to almost everywhere in the world. The city attracts a large volume of diverse talent, from which the University and the Department have benefited and will continue to benefit when recruitingfaculty and students. Third, the Department has received sizable fundsfrom the College to support graduate teaching and research assistants. This has proved crucial to this new Department, considering the currentlytight Federal funding situation for academic research. Fourth, the University provides tuition waivers to funded research assistants, which makes its graduate programs cost-effective and competitive. Fifth, a strong focus on bioinformatics is giving the Department a unique, national identity. This emphasis has also effectively leveraged the strengths of other departments in the college, such as Biology and Chemistry. Finally, in multiple small but cumulatively significant ways, Central Administration has backed their words of support for the Department with concrete,positive actions.
The CSD has strong potential to move to the ``next level’’of national recognition, and it is toward achieving this goal that this evaluation document is addressed. To attain this goal, it is important to identify opportunities for improvement and takeprompt action; this is particularly important to a young, growing department. These opportunities are listed below. Specific recommendations arefound insubsequent sections of this report.
The undergraduate teaching structure, curriculum, and advising/tutoring environment should be upgraded. Currently, the lower-level undergraduate classes are taughtalmost entirely by instructors and graduate students,negatively impacting quality. Faculty with Ph.D. degrees, such as lecturers and tenure-track and tenured faculty (TT faculty), should be committed tolower-level undergraduate teaching. Next, the undergraduate curriculum possessesa traditional but dated structure and requires revision to encompassrecent technology changes. Finally, the department does not possess a formal advising and tutoring service for undergraduate students; sucha service is necessary to bolsterstudent retention, course and curriculum planning, and aidstudents with career planning.
The current space available to the Department is woefully inadequate for education and research activities. The Department urgently needs office space for its graduate students, especially thosewho work as teaching and research assistants, and there is also a desperate need for laboratory space for thosefaculty who conduct experimental research. Also lacking ismeeting, study, and social space for the Department’s undergraduate students, who are scattered across the cityand have no ready means of interacting with faculty and with one another.
Faculty research qualitymust be more strongly emphasized. A department’s reputation is generated fromthe quality andimpact of its faculty’s research, not frommerequantity of publications. The Departmenthas been prolific in publishing papers, and in order to improve its reputation, visibility, and national ranking, the Department must now increase its emphasis on quality: Faculty musttarget those conferences and journals that make highest impact in their fields. One drawback of the Department’s current faculty composition is that allfull professors are relatively junior; the Department would benefit from a senior full professor who canlead the way in research quality.
Guidance of junior-faculty career development should be undertakenmore systematically. Our interviews with junior faculty members led us to conclude that the junior faculty lack precise information about the Department’s tenure and promotion process and expectations. Although the University has stated procedures and expectations, the Department mustclarify what these mean within the CSD. Issues that must be clarified for the tenure review include: research productivity in terms of quantity and quality; the required quantity, quality, and level of teaching; required service work; the process of selecting external reviewers for tenure review and the format of the external reviews; and the purpose and format of the mid-probationary review. The precise process of the tenure review itself should alsobe clearly stated.
Quality of the graduate program should be made highest priority. The Department has been truly successful atPh.D. production. Current student numbers appear to be at or near a maximum level for the current size of the faculty. Looking towards the future, the Department should plan for Ph.D. production between 5 and 7 per year, where quality of the dissertations takes priority over quantity. In addition, faculty should begin to pay attention to optimal placement of their graduating students, because well-placed graduates accelerate the visibility and prestige of the Department.
Staff support is insufficient. The department currently has 4 administrative staff and one IT staff person --- this appears to be insufficient. A full-time staff person is needed for administrative support for graduate admission and management, and a secondIT staff member is needed to support the large number of experimentally oriented faculty and student research projects.
Faculty recruiting must be based on a strategic vision. Because the CSD is small, each new faculty hire is critical. The Department should conceive a strategic vision of how to achieveits desired national reputation and use this vision to guide its recruiting. Although it is tempting to recruit faculty who have already acquired research funding, this factor in itself weighs less than hiring someone who helps fulfill the Department’s vision. The Department should avoid building towards a research area that is already crowded nationally and select instead achievable ``niche areas’’ that build upon the Department’s existing strengths.
- Historical and Current Context
Given the short history of the Department, the undergraduate and graduate programs are very productive and well functioning. The research productivity of the faculty is adequate for its graduate program. The Department is an important educational asset to the city of Atlanta, providing both undergraduate and graduate studies, and the regional demand for graduates of its programs seems to be high. The peer institutions selected by the Department, such as GeorgeMasonUniversity and SouthFloridaUniversity, are appropriate.
C. Progress toward Goals and Objectives
This is the Department’s first external review.
The Department has succeeded in its initial goals to serve a large number of undergraduate students and also to generate a sizable graduate program within a relatively short period of time and with limited resources. The recent decline in its undergraduate enrollments has tracked the overall national decline in CS enrollment. It is anticipated that CS enrollment will now grow again. The next stage in the Department’s development should be to maintainmanaged growth andimprove further the quality of both undergraduate and graduate programs.
- Quality of the Curriculum
Despite the fluctuation in enrollment numbers, as noted in the tableon Page 15 of the Department's Self Study document, the Department'sBachelor of Science program retains a healthy enrollment; following anticipated national trends, the number of undergraduates will increase over the next 5 years. The large number of HOPE scholarships made available by the State ofGeorgia to the Department's studentshas improved student quality and makes possible a high-quality undergraduate program.
As noted in Tables 5a-5c, Page 11, of theSelf Study document, the faculty's commitment to buildingan attractive graduate program has required thattemporary instructors and graduate students teach virtually all ofthe Department's lower-level (required freshman and sophomore) courses. As an example, during the 2006-07 academic year, 22of the 25 sections of the Department’s lower-level courses were taught by temporary instructors. The Department's retention rate of undergraduates is lower thanthe University's average; this seems partially due tothe utilization of the temporary teachers as well as the absence of a departmental advisor for course and curriculumplanning. These are weaknesses that must be addressed.
The structure of the B.S.curriculum is rigid,making it difficult for students to complete in 4 years. Finally, given the large number of required physics, mathematics, and CS courses, students have no opportunity to undertake dual majors, major-minor arrangements, or multidisciplinary studies.This is in stark contrast tothe Department'sgraduate curriculum, whichisquiteinterdisciplinary in nature (e.g., bioinformatics) and highly successful.
Students' progress towards their degrees also ranks below theUniversity's average; this appears to be partlyconnected to the curriculum. The Department's Bachelor of Science curriculum requires
- 21 credits of mathematics, 12 credits of physics,
- 25 credits of lower-level computer science,
- 16 credits of upper-level computer science, and
- 6 credits of advanced computing-related electives.
It is worth contrasting these requirements to those of the Computer Science Accreditation Board of the Accreditation Board for Engineering and Technology (ABET)for a computer-science curriculum: ABET requires (please note the differences in italics)
- 15 credits of mathematics, 12 credits of natural sciences,
- 16 credits of lower-level computer science,
- 16 credits of upper-level computer science,
- 8credits of advanced computing-related electives.
The differences arestriking and suggest that the existing curriculum hampers the Department’s students fromprogressingtowards upper-division courses and limitsthe number of courses a student may take to develop a specialization. A revision of the B.S. curriculum, consistent with ABET guidelines, would organize the curriculum into modules:
- a lower-level module of core computing courses that also would constitute a minor-degree emphasis for students undertaking interdisciplinary studies
- an upper-level module of advanced computing courses
- a specialization module that would allow and encourage cutting-edge or interdisciplinary studies.
The proposed lower-level module would be slightly faster paced that the existing curriculum’s, but interviews with the Department’s undergraduates suggest that this would not be a significant obstacle.
In addition, ABET’s acceptance of natural science courses beyond physics opens the doorto students to undertake dual-major and major-minor studies in areas such as biology and chemistry, where the Department has already established interdisciplinary programs at the graduate level. From here, it would be a small step to integrate undergraduates intothe Department's interdisciplinaryresearch projects and generate amongst the undergraduates the sameexcitement that is now found in the Department's graduate students.
A concern stated earlier is the Department's employment oftemporary instructors and graduate students to teach required lower-level courses. Student input indicates clearly that significant variations inquality have arisen between different sections of the same coursedue to varying quality of the temporary teaching staff. Further,the lack of contact between TT faculty and beginningstudents in these key, core courses has negatively impactedstudent satisfaction, has lowered student retention rates,and will ultimately affect quantitative measures used by organizationslike U.S. News to rank undergraduate programs.
Given that the Department must maintain current instruction andresearch standards of itsupper-level undergraduate and graduate programs, one concludes that 3-4 additional Ph.D.-level facultyare required to manage, teach, and maintain the four core, multi-section, lower-level, computer science courses.
The current graduate-level curriculum is well designedand works well for the Department. The graduate students were particularly appreciative of the flexibility recently introduced into the qualifying exam structure.
E. Quality of the Students
GSU has done a commendable job at improving thequality of its incoming students. These efforts, paired with theHOPE scholarships, have positively impacted the quality of studentwho pursues undergraduate computer-science studies. The match ofstudent to level of curriculum appears to be appropriate.
Several faculty stated concerns regarding recent selectivity of graduate students. An obvious shortcoming is the low stipend paid to graduate studentswho are supported as TAs ($12,000 per year, where the national average runs closer to $18,000 per year). Increasing the level of stipends is one step forwards. The support level for Fellows and grant-supported RAs is competitive.
F. Quality of the Faculty
During the interview session, the undergraduate students indicatedsatisfaction with both the teaching ability and the intellectuallevel exhibited by the tenure-track faculty. The faculty members were alsocommended for their willingness to meet with students,even when subjected to severe time pressures. The graduate students who were interviewed expressed similar sentiments.
Most faculty members in the department are actively engaged in research with continuous publications annually. The Department’s research culture has improved as new, research-active faculty at both senior and junior levels have been hired in recent years. A major challenge facing the Departmentis the placement of their Ph.D. graduates into jobs at competitive research institutions ---a few high-profile Ph.D.-graduate placements might well be all that is needed to place the Computer Science Department ``on the map’’ of the academic research community.
In the area of research publications, the faculty members should select well-respected venues in which to publish their research results. The number of publications is much less important than the quality of the papers, and we encourage the faculty to make their best efforts to aim at publishing their work in the leading venues in their research areas.
G. Resource Adequacy
To move to the next level of quality and recognition, the department needs to increase its resources by increasing the number of TT faculty for both graduate and undergraduate programs, by adding space for graduate and undergraduate students,research and teaching labs and conferencerooms, and by increasing administrative and technical staff support.
Undergraduate students expressed some confusionregarding the computing and lab facilities available for their use.It would be helpful to the undergraduates if a clear descriptionof computational and other resources available to them is provided.
Although many undergraduates own their own computers, those students who do not have been performing their work with the assistance of University-managed computing labs.
Recent purchases of digital subscriptions to the main-line ACM and IEEEpublications plus the nearby presence of the GeorgiaTechUniversity librarymake the library facilities required for computing research adequate.
H. Goals and Objectives
Following is a listing of the Department’s goals from its Self-Study Document, editedto retain those goals we consider most important.
Goal 1. Increasing the number B.S. graduates by increasing the number of undergraduate students and improving quality and retention
Recommendation:
A strong undergraduate program provides a foundation for a department's graduate programs --- no high-quality graduate computing program lacks a similar high-quality undergraduate program. A strong undergraduate program provides a ``pipeline’’ that leads talented students into graduate study.
The Department’s goals can be facilitated by a redesign of the undergraduateB.S. curriculum so that it matches the ABET requirements.This will relax the curriculum’s restrictive structure andencourage interdisciplinarystudy, a Departmental strength. During the redesign process, consideration should be givento implementing cooperativedegree programs with Biology andChemistry, exploiting the already successful collaborations at thegraduate level.
Quality and retention can be addressed by the infusion of 3 to 4 additionalPh.D.-level faculty (a mix of lecturers and TT faculty) into the undergraduate teaching staff.The existing Instructor positions should be converted into Ph.D.-levelLecturer and Assistant Professor positions, and new TT facultylines should be added. Accompanying the curriculum revisions should be an action plan that describeshow TTfaculty will be utilizedto manage and maintain the lower-level, undergraduate courses.