5
A Positive Environment for Academic ECSE

Previous chapters have discussed infrastructure, research evaluation, and educational issues as they relate to experimental computer scientists and engineers in academia. The committee intends this chapter to be a description of the characteristics of a positive academic environment in which experimental computer science and engineering (ECSE) can flourish, rather than criticism of any individual department. Recognizing that ECSE is thriving in some departments, but not in others, the committee expressly chose this formulation as a means of encapsulating those characteristics that seem to work well.

MENTORING AND ADVOCACY

When Odysseus set out for Troy, he entrusted the care of his household to Mentor. Although the burden to succeed properly belongs to the researcher, young researchers setting out on their academic careers still need mentors, loyal friends, wise advisers, teachers, faithful counselors, guardians, and advocates to help advance their interests. Young experimentalists often face greater and more complicated demands than do their theoretically inclined colleagues, many of which follow from the project-oriented nature of the work.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 82
Academic Careers for Experimental Computer Scientists and Engineers 5 A Positive Environment for Academic ECSE Previous chapters have discussed infrastructure, research evaluation, and educational issues as they relate to experimental computer scientists and engineers in academia. The committee intends this chapter to be a description of the characteristics of a positive academic environment in which experimental computer science and engineering (ECSE) can flourish, rather than criticism of any individual department. Recognizing that ECSE is thriving in some departments, but not in others, the committee expressly chose this formulation as a means of encapsulating those characteristics that seem to work well. MENTORING AND ADVOCACY When Odysseus set out for Troy, he entrusted the care of his household to Mentor. Although the burden to succeed properly belongs to the researcher, young researchers setting out on their academic careers still need mentors, loyal friends, wise advisers, teachers, faithful counselors, guardians, and advocates to help advance their interests. Young experimentalists often face greater and more complicated demands than do their theoretically inclined colleagues, many of which follow from the project-oriented nature of the work.

OCR for page 82
Academic Careers for Experimental Computer Scientists and Engineers Mentoring Recognizing these differences, computer science and engineering departments need to be proactive in helping to establish mentoring and advocacy relationships between junior and senior faculty members. (As used here, mentoring refers to advice from a senior faculty member to a junior faculty member. Advocacy refers to commentary, input, and argument from a senior faculty member to department chairs, deans, and others higher up in the university hierarchy in arguing for and promoting the interests of a junior faculty member.) These arrangements should be made openly and explicitly. Young faculty should know from the start whom to consult for advice and counseling. They will appreciate the help and attention, and the department will have created a positive factor in retaining and recruiting first-rate faculty. The senior faculty members who take on mentoring and advocacy roles should be encouraged to meet, discuss their situations, and find ways to support each other. The more general types of guidance are listed below, and Box 5.1 describes specific mentoring tasks. Given the time and resource demands of ECSE, a junior faculty member must "hit the ground running" to be successful. In all but the most unusual cases, the process is bootstrapped: early research success using start-up resources and no graduate students is parlayed into funding that can support a more ambitious implementation effort with graduate students who should by then be trained. The more ambitious artifact must be completed in time to perform the experiments so that the results can be disseminated to the community early enough for the work to be evaluated by the tenure letter writers. Senior faculty mentors have an important role in facilitating such an outcome. In addition to technical assistance, senior faculty can provide advice about practical aspects of experimental work, such as managing time, money, and space. They can also provide guidance about the expectations for tenure. Following are several areas in which mentoring senior faculty can play important roles: Establishing cooperative and collaborative environments. Since junior faculty generally lack reputations that attract resources, they are often dependent on senior faculty to obtain entry to established laboratories that can provide needed equipment, staff, and technical skills, as well as an intellectual community. Junior faculty without funding of their own can participate in existing grants while seeking independent sources of funding. New projects that start surrounded by established activities enjoy an increased likelihood of success.

OCR for page 82
Academic Careers for Experimental Computer Scientists and Engineers BOX 5.1 A Mentoring Checklist Publication In which journals and conferences should the junior faculty member publish? How often? Can the mentor review papers before submission for publication? Funding Which funding agencies support the type of work the junior faculty member wants to do? Which industrial companies share the junior faculty member's intellectual interests? Who within agencies and companies are the right people to meet? To what extent are university start-up funds adequate for a new junior faculty member's initial work? Can the mentor review the junior faculty member's grant proposals? Collaboration What senior people in the field share intellectual interests with the junior faculty member? Who has laboratories or other resources that are shareable with the junior faculty member? Visibility Who are the senior people in the field who should know the junior faculty member's work? How can seminars or other presentation forums be arranged to showcase the junior faculty member's work? What contributions has the junior faculty member made that are not widely recognized? Problem Choice What problems should the junior faculty member choose that are doable in the given academic environment and have the potential for substantial impact on the field? How can a research program be structured so that it has meaningful intermediate outputs? What are dead-end problems that should be avoided? Is the project consistent with the resources that will be available? Students and Teaching How can good and appropriate students be attracted to work with junior faculty? How can the junior faculty member's teaching be improved? At what point is the junior faculty member spending too much time on teaching and education?

OCR for page 82
Academic Careers for Experimental Computer Scientists and Engineers Service What service activities should the junior faculty member perform or avoid? How can the junior faculty member be protected from inappropriate service work? Logistics What vendors should supply the junior faculty member's equipment or software? Who in the university should be approached about obtaining laboratory space? Matching project scale and scope to available resources. Because they lack experience in research management, junior faculty are often unable to assess the appropriate scale of a project relative to their operating environment. In their enthusiasm, junior faculty members may embark on projects that are too large or complex given the resources likely to be available. Senior faculty can give advice about what is reasonable given the limits of available resources. Later, as an advocate to the department and the university, the mentor may be called on to explain how the young researcher's accomplishments are consonant with the resources at hand. Improving the visibility of protégés. Senior faculty can play a key role in generating speaking opportunities for junior faculty at departmental colloquia, industrial research laboratories, or other universities, workshops, and other settings in which the junior faculty member's work can be showcased. Perhaps more importantly, they can also encourage junior faculty to make such presentations. Recognizing collaborative contributions. The collaborative nature of many ECSE research projects is often at odds with the need of the tenure and promotion (T&P) process to identify contributions made by specific individuals. Such identification may be particularly important in the case of an individual who makes intellectually substantive contributions to an unsuccessful project that failed for entirely separate reasons. Given that collaborative research projects are most often carried out under the direction of a senior faculty member, junior faculty collaborators may not be recognized publicly for their specific contributions without explicit acknowledgment and promotion of their efforts by the senior project director. Counseling protégés to adopt research strategies that produce significant intermediate results. Even in acknowledging the fact that ECSE research may involve "all-or-nothing" risks, researchers are still best

OCR for page 82
Academic Careers for Experimental Computer Scientists and Engineers advised to avoid mega-projects that after years of producing nothing emit a single, definitive magnum opus. A more conservative approach, which often requires the guidance of an experienced researcher, is to structure the study so that it produces intermediate results suitable for publication, thereby balancing significance and scope against the need for visible output. Examples might include reporting results from simulations used in designing the artifact, describing technology used to build the artifact, or presenting analytical studies of some aspect of the artifact. Intermediate publications serve as concrete evidence of research progress. Additionally, they provide early exposure of the ideas to the community that can result in prompt feedback useful for midcourse corrections. As the discussion above suggests, a mentor must be intimately familiar with the ECSE field, research practices, and community. It is important for mentors to understand the career and research goals of their ostensible protégés. The senior faculty have to take responsibility for becoming familiar with the peer groups, conferences, and organizations that are important to a young researcher's career. Wellmeaning though a senior theoretician may be, the mentoring role for a junior experimentalist is best served by a senior experimentalist. A department that does not have appropriate senior experimental faculty to serve as mentors for junior faculty should consider finding someone at another university or laboratory who can play the role of outside adviser, perhaps by using regular visits (at least once or twice a year), telephone calls, and e-mail to maintain frequent contact with the young faculty. (A new faculty member's doctoral adviser is ideally situated to play such a role.) External mentors can play the same sort of role that visiting committees play during the evaluation of departments—providing advice, offering contacts, and being a sounding board for new ideas. This is also a way for universities without large, established experimental programs to develop such programs. Advocacy Although the roles of advocates and mentors overlap, they are somewhat different. Whereas a mentor gives advice and counsel and guides the junior faculty member through the academic jungle, an advocate is distinctly partisan. The advocate's job is to advance the interests of junior faculty members and to help them make the best possible case for their promotion. Perhaps the most important role of an advocate is laying the groundwork for a tenure case within the university. University-wide (or

OCR for page 82
Academic Careers for Experimental Computer Scientists and Engineers BOX 5.2 An Advocate's Checklist The functions of an advocate in the tenure and promotion process include: Helping to select or nominate specific letter writers familiar with the candidate's work (including collaborators), based on their ability to evaluate such work; Explaining and documenting key characteristics of ECSE to those outside the field, pointing out potential mismatches between ECSE and more traditional disciplines; Accumulating evidence of impact of the junior faculty member's work on practice, including letters from industrial contacts, and arguing the case that industrial scientists are well-placed to judge impact; Obtaining letters from other researchers who may have used parts of the experiments in their own work, or who may have used the experimental results to guide their work; Soliciting referee reports from conferences to document the quality of conference contributions; Documenting how negative experimental results may have helped drive the field in a positive way; Explaining the structure and pecking order of the literature; knowing which conferences and journals are respected, prestigious, and well refereed; being knowledgeable about acceptance rates and review procedures for conferences and private journals in which the junior faculty member has made presentations or published; Explaining the significance of artifacts produced by the junior faculty member; Understanding historical matches between resources and project scale; Explaining the importance of collaborative work in ECSE; and Extrapolating the future performance of the junior faculty member. even school-wide) T&P committees may not understand the nature of ECSE very well and may attempt to judge tenure candidates according to inappropriate criteria. Accordingly, an effective advocate has to know the field well—its standards, its interesting questions, its history, its characteristic work modes—and be able to communicate the goals and aspirations of ECSE as a discipline to others. Box 5.2 describes some of the things that advocates may have to do to support the tenure case of a junior faculty member. A forceful and experienced advocate may be particularly important in those universities that have not institutionally recognized the

OCR for page 82
Academic Careers for Experimental Computer Scientists and Engineers significance of artifacts in scholarly endeavors. At these universities, an advocate may have to argue anew for every ECSE promotion or tenure case that the creation of artifacts can be a legitimate focus of scholarly research. DESIDERATA FOR THE TENURE AND PROMOTION PROCESS Universities evaluating candidates for tenure or promotion take into account a number of indicators. The committee recognizes a wide range of approaches to evaluating candidates for tenure and promotion, and it does not wish to intrude on institutional prerogatives in determining how best to evaluate candidates. At the same time, the committee believes that evaluators should use standards and criteria that normally characterize productive work in the ECSE discipline, rather than standards that may be applicable to more traditional academic disciplines. Care should be taken not to exclude meaningful evidence of achievement simply because it is nonstandard. Indeed, the committee believes that T&P committees and university administrators should take a catholic perspective on the available evidence, regardless of the different forms in which such evidence appears. The purpose of the discussion below is to point out how characteristics of the ECSE discipline may affect the indicators that universities take into account in making T&P decisions and how those indicators should be evaluated. However, it is clearly the prerogative of individual universities to determine the relative weight that each indicator should carry for T&P candidates. The committee observes that some institutions, notably those with a strong and continuing tradition of experimental work, already take these characteristics into account in their T&P processes. Publications Many universities regard archival journal and book publications as the primary medium in which scholarly work is demonstrated, whereas ''mere" presentations at professional meetings are regarded as second-rate. Although such practices are not the rule at universities with strong experimental traditions, anecdotal evidence suggests that they are considerably more common in universities without such traditions. As noted in Chapter 4, publications in ECSE take a variety of forms: technical reports, conference proceedings, or articles in archi-

OCR for page 82
Academic Careers for Experimental Computer Scientists and Engineers val journals. For ECSE, publication in certain conference proceedings may carry as much or more weight than publication in highly regarded journals. The dilemma of choosing between the requirements for the "proper form" of an academic record and the "content" for ECSE was nicely stated by an assistant professor at a major private university in response to the CRA-CSTB survey: Tenure means that I have to spend an enormous amount of time writing papers for archival journals and conferences, so that people can peer-review me without understanding what I do. Most of the impact of [experimental] research work comes from dissemination channels such as e-mail, via which the software artifacts produced by the research can be spread into the community. Given this tension, candidates for promotion in ECSE may face a significant disadvantage compared to their more theoretically inclined colleagues. Specifically, their publication portfolios may well be shorter (because of the time-consuming nature of ECSE research) and may contain fewer publications in archival journals (because of the field's preference for the timeliness of conference publication). Such characteristics should not prejudice a candidate's case, if it is documented in other ways that important and recognized intellectual output has been produced (e.g., through the production of significant computing artifacts, as described below). ECSE articles in archival journals can be expected primarily at the end of a project, independent of whether the experiment was a success or failure. Such articles distill project results and summarize the issues raised, the insights gained, and the implications for future research. They unify the results of the project and describe how it fits within the larger context of the field. Accordingly, it is not uncommon to find a single archival journal publication for an entire multiyear project. By contrast, conference publications relate work in progress and intermediate milestones. A typical project may result in several conference publications but only one journal publication. Technical reports complement both conference and journal publications by describing the project in substantial detail; such reports are an essential vehicle for disseminating the specifics of the project. The timing of journal publications as it relates to the nature of the discipline should also be recognized. It is not uncommon for a T&P committee to interpret an unevenly distributed publication record (e.g., a long gap with no publications followed by a "surge" of publications near the tenure decision) as a response to the tenure decision rather than as an indication of true productivity. In the case of experimentalists, this interpretation should be examined carefully, be-

OCR for page 82
Academic Careers for Experimental Computer Scientists and Engineers cause the time scales of project completion are often comparable to the probationary period. T&P committees might wish to examine the extent to which the publications taken together indicate a coherent intellectual theme. If so, a publication surge is less suspect than if the publications are relatively disconnected. Faculties, deans, and university administrators generally disclaim in public assertions that simplistic methods such as publication counting are used in considering candidates for tenure. Whether or not such methods are used at any given university is difficult to determine, although there is a widespread perception that publication counting is widely practiced. It would not be surprising if it were practiced at many universities, because the appropriate basis for a promotion decision—significance and impact of research—is difficult to verify independently by committees far removed from the candidate's research specialty. In short, a candidate's record of publication in archival journals is only one aspect of the individual's overall portfolio, and for ECSE perhaps a misleading one at that. The candidate's record in producing innovative and useful artifacts of high quality and the letters of recommendation supporting the promotion may be better indicators of his or her history and likely future performance. In any event, publication portfolios should include documentation regarding matters such as frequency of publication, acceptance rates, and publishing history for journal or conference publications, as well as the board of editors or program committee members. These can help evaluation committees to understand the basis used to determine the worthiness of those publications. Artifacts A track record that might appear modest when assessed by counting journal articles may in fact be truly spectacular when evaluated in the context of a discipline in which a technical reputation is founded as much on functional artifacts as it is on publications. Production of artifacts is so important to the field that a standard part of any experimentalist's curriculum vitae should be a section describing computing artifacts produced by the experimentalist. However, the reality in many universities—especially those without strong engineering traditions—is that artifacts are mentioned only in the context of such ''creative fields" as music, art, and theater; evaluations are conducted on the basis of the quality of the artistic production and make use of evidence such as published reviews of performances or awards in juried exhibitions. Thus, the importance of artifacts to ECSE research may need to be clearly established as a principle. Once the principle has been established, the focus should

OCR for page 82
Academic Careers for Experimental Computer Scientists and Engineers turn to the impact of the artifacts that the candidate has produced. One obvious dimension is intellectual impact—to what extent has an artifact had an impact on researchers in the field? This question is treated in depth in Chapter 1. However, another dimension of impact—impact on practice—is, in the committee's view, underappreciated. An ECSE researcher who creates an innovative computing artifact whose primary impact has been on practice (e.g., an artifact that is embodied in a large number of nonresearch systems or used by a large number of "just plain users") has made a substantive and meaningful contribution in the tradition of ECSE, similar to that of a theoretician who proves an important lemma. Of course, simplistic measures such as "number of users to whom the artifact has been disseminated" or "number of FTP downloads of software from the researcher's laboratory" are as meaningless as publication counts. The true impact of an artifact is documented more meaningfully by letters from prominent and trustworthy colleagues in academia or industry, and any other users of the artifact. In addition, the scale of the project that produces any given artifact should be an important consideration in its evaluation. Specifically, it is inappropriate to expect a small-scale experimental project (e.g., one on the order of $100,000) to produce results that are comparable to those attained by projects 10 times that size. Review Letters Letters that evaluate the accomplishments and the promise of candidates for promotion (especially at the point of tenure) are an integral part of the candidate's portfolio. Indeed, for academicians in ECSE, as in other disciplines, letters may be the most important component of the portfolio. The primary reason is simple: impact on and value to others are the key qualities to be established in an individual's work. Honest and well-documented letters by knowledgeable evaluators are the best way to demonstrate impact and testify to value. A secondary reason is that letters expressing the judgment of senior researchers may be necessary to identify ideas or artifacts with large but as-yet unrealized potential for impact.1 1   The long time scales required for artifact implementation may well mean that a good idea does not have time to diffuse into the community at large during the first six years of an assistant professor's career. Moreover, the difference between a project that demonstrates the technical feasibility of a promising concept and one that develops the proof-of-concept prototype to releasable form may be a factor of 10 in resources spent, even if no new ideas emerge as those additional resources are spent. In other cases, a useful innovation may be part of a larger system that will be deployed in the future, thus restricting experience with the actual implementation.

OCR for page 82
Academic Careers for Experimental Computer Scientists and Engineers It is important that letter writers be provided with enough information to make valid and useful judgments, such as copies of the papers highlighted by the candidate in the publication section of the curriculum vitae and obviously a copy of the curriculum vitae itself. The reviewer should make explicit comparisons with other faculty members in the candidate's peer group. Given the importance of letters, the selection of letter writers becomes a critical problem. The committee believes that the primary criteria in selecting potential letter writers should be their stature in the field and their familiarity with the candidate's work. Other factors, such as the letter writer's institutional location or status as a research collaborator of the candidate, should not be reasons for eschewing letters from such individuals. Of course, good arguments can be made to support the proposition that letter writers should not consist exclusively of collaborators or industrial scientists because of potential bias, unfamiliarity with the academic environment, and so on. However, to exclude letters from such individuals or to impose arbitrary limits on these letters is as inappropriate as including only letters written by senior academics who have no personal knowledge of the candidate's work. It is particularly important that letters from individuals in industry not be limited arbitrarily; such letters should carry a weight equal to those of similarly qualified and reputable individuals in academia. The reason is that ECSE work with high impact is likely to affect industry. Much of the most advanced work in ECSE is done in industry, and many of the top researchers in ECSE are found there. (Analogous remarks apply to qualified and reputable ECSE researchers at government laboratories.) Similarly, the potential evaluator's reputation within the field, and his or her knowledge of the candidate's work, are far more important than the overall reputation of the evaluator's home institution. Although this proposition may seem obvious, the committee found many examples of universities in which review letters from industry scientists and engineers are not only discouraged but often never sought. As for collaborators, the concern that they may be unduly biased toward (or, on occasion, against) the candidate on the basis of existing personal relationship is a valid one. However, to avoid letters from collaborators in a field as intrinsically collaborative as ECSE is to eliminate some of the best possible input regarding the candidate's intellectual capacity, creativity, and originality. Documenting the extent and nature of an individual's contributions is surely required. The best way to find out, of course, is to ask the principals and read their letters with care. Attempts to allocate specific percentages of credit to individuals for collaborative work are foolish in the extreme.

OCR for page 82
Academic Careers for Experimental Computer Scientists and Engineers Funding History Tenure and promotion committees often take into account the candidate's track record in obtaining research funding, on the assumption that the ability to attract outside funding is an indicator of competence. Often, faculty members are evaluated on the aggressiveness with which they have sought out available research opportunities and how effectively they have met the expectations of funding agencies. Most ECSE researchers do require considerable funding in order to pursue their research; the only exceptions are those instances in which the faculty member is fortunate enough to make connections with an existing laboratory and experimental infrastructure or those in which the faculty member has been able to develop an experimental research program "on the cheap." Although funding is an enabling factor for ECSE research, it is not in itself necessarily demonstrative of intellectual achievement. Indeed, a faculty member who has structured his or her research to produce meaningful results, with high impact, on a limited budget deserves praise for creativity and good problem selection, rather than censure for not producing dollars for the university. Government funding decisions—even at strongly peer-reviewed agencies such as the National Science Foundation—are not (and should not be) based simply on a rank ordering of proposal quality and a cutoff above some specified line. Although reviewer scores of proposals could be used to generate a rank ordering of those proposals, program directors are expected to take other considerations into account. "Hot" topics wax and wane, and reviewer scores tend to trail the curve—so some topics are often scored higher simply because the reviewers are more familiar with them. One job of a good program director is to identify new topics that could extend the frontiers of the field in novel directions, which sometimes means funding proposals in new areas that may not have received particularly good scores. A second job of program directors is to maintain a good balance of topics in their portfolios—so slightly poorer proposals in an emerging area ought to be chosen ahead of slightly better ones in oversubscribed areas. In short, program directors who do their jobs well may not be funding only the highly meritorious proposals. Consequently, a paucity of research funding should not be held against the junior faculty member who has otherwise demonstrated an adequate level of significant research productivity. As noted above, industry can be an important source for the funding of ECSE research. Industry funds can support equipment grants,

OCR for page 82
Academic Careers for Experimental Computer Scientists and Engineers fellowships for students, small cash grants, and major cooperative efforts. In many cases, especially with equipment grants, industry decisions seem more likely to be made on the basis of the reputation of the school than of the individual. Nevertheless, individuals deserve credit for receiving this kind of support. Cooperative research and development grants are sometimes problematic, because much of the work may be done at the company and relatively little money given to the university. Therefore, what may look like a minor project on the résumé may actually constitute a significant achievement with major impact. Reviewing industrial grants is normally done strictly within a company, but obtaining such a grant is generally a reliable indication of impact when a long-term relation can be established. Other considerations in the T&P Process Ph.D. Students Given the nationwide average of 6.4 years for students to complete Ph.D. degrees, it is unrealistic to expect a junior ECSE faculty member to produce very many Ph.D. students before a tenure decision must be made; he or she may be fortunate to have graduated one, especially if a research laboratory and team had to be built from scratch. However, it is not unreasonable to expect an assistant professor at this time to have a number of students in the pipeline and one or more close to graduation. More important than the number of graduated Ph.D. students at this stage is their intellectual and professional development. Relevant indicators may include their production of useful and novel artifacts (even if small scale), their intellectual independence, and even their records of public presentations of work. Moreover, because incoming graduate students may tend to avoid junior faculty in favor of senior faculty with established reputations, such assessment should also take note of the quality of the students available to the junior faculty member. Consulting Consulting can be an indicator of the research impact of a candidate's work. However, a full description of consulting work (including products that have been or may be developed under the consulting arrangement) is necessary for T&P committees to judge this impact. Nondisclosure agreements are thus particularly problematic for

OCR for page 82
Academic Careers for Experimental Computer Scientists and Engineers junior faculty members. Given the tradition of academic research as work that may be freely published and disseminated, consultants whose primary allegiances are to academia are urged to resist restrictive nondisclosure agreements to the maximum extent possible. Consulting relationships that focus on implementation at the expense of reportable scholarly work may remunerate the faculty member/consultant but do not advance the research enterprise. Instead, faculty members should seek consulting relationships that strengthen or enhance their own research programs or that provide opportunities, such as scholarly publication, in which they can obtain public recognition for their work. For example, faculty members may be able to negotiate short and finite periods of time in which they will refrain from discussing their work in public, or they may agree that technical details should be kept private but conceptual ideas can be freely discussed. Patent applications might be considered. However, in the absence of such expressions of understanding, faculty members who undertake proprietary consulting work do so at their own risk with respect to the T&P process. A second potentially negative aspect of consulting should be considered as well. Given the financial rewards available from consulting for industry, ECSE faculty, especially those with tenure, have great incentives to undertake such activities. If too much time is spent on consulting activities, the perception of a less-than-dedicated "part-timer" occupying a full-time faculty slot may take hold. Such perceptions may be rather negative, especially in publicly funded institutions whose legislatures may already be concerned about "faculty who spend so little time teaching." Still, despite the potential pitfalls, consulting for industry can carry a variety of benefits for faculty members entirely apart from the remuneration involved. It can familiarize them with the problems facing industry, thereby suggesting potentially interesting research directions. It can help to place a faculty member's teaching in a better context, especially for students who will eventually work in industry (see discussion on teaching below). Finally, it can be the basis for acquiring additional research funding. Workshops Workshops are specialized conferences with a focused theme. In many cases, participation is by invitation only; in other cases, the program committee selects from extended abstracts submitted by researchers in the area. Workshops play an important role in setting research directions in a given area and the timely exchange of ideas.

OCR for page 82
Academic Careers for Experimental Computer Scientists and Engineers In this way, they can be superior to conferences in other fields. Participation in workshops can demonstrate community acceptance of a researcher's competence if not the researcher's work per se. Teaching2 The rapid changes in the technological substrate of ECSE mean that even lower-division ECSE courses will need frequent updating, much more so than courses (e.g., in the physical sciences) in which the fundamentals remain the same from year to year. In addition, recent reports such as Computing the Future have called for the development of interdisciplinary connections between computer science and other problem domains, and the development and teaching of interdisciplinary courses would be evidence that the individuals involved were furthering the interests of the field. Rapid change affects the core curriculum of CS&E even at the graduate level. This provides significant opportunities for junior faculty to develop new courses in or near the area of their research—opportunities that may be less available in more established science and engineering disciplines such as physics, whose core courses tend to change much more slowly. At the same time, exploiting such 2   The relatively brief discussion of teaching in this section should not be taken to imply a committee judgment about the relative importance of teaching versus research. The issues raised in this section are issues that are characteristic of ECSE vis-à-vis other disciplines, but the much broader question of the appropriate balance between teaching and research is outside the scope of this report, as are questions such as, How much credit should a candidate for tenure or promotion receive for the writing of a textbook versus obtaining excellent student evaluations? Nevertheless, a number of the faculty who responded to the CRA-CSTB survey made spontaneous comments regarding the teaching-research balance. Here are two: I find that the pressure to publish and get money detracts from the undergraduates. They are often short-changed, and I believe that the situation cannot continue. The nation as a whole is not being served well, if the only people who get tenure are those who have little interest in teaching undergraduates. Is our nation's policy of ignoring undergraduates a significant reason why the majority of our graduate students come from overseas? Keep your teaching and administrative responsibilities in perspective—if your research suffers too much, then you won't get tenure. Do a reasonable job at teaching and do carry out essential administrative chores, and most importantly be sensitive to your [own] students' needs, but remember this: there is no end to the distractions that will keep you from doing your research and writing those papers. You will have to say no at times when you would rather not, but if you don't learn how to say no when it's necessary (and to do so nicely), then your life will become very, very difficult.

OCR for page 82
Academic Careers for Experimental Computer Scientists and Engineers opportunities is often very time-consuming for junior ECSE faculty, because there is often no suitable textbook even for core courses in ECSE. Advocates for junior faculty will want to ensure that the significance of such course development does not go unnoticed at promotion time. The competent supervision of undergraduate research in ECSE is a notable accomplishment in a teaching portfolio, especially given the difficulties in formulating a meaningful ECSE research problem that an undergraduate can plausibly undertake. Undergraduate papers or presentations reporting research results would be a powerful indicator of such supervision. Teaching has an industrial dimension as well. Terms such as partnerships with industry have entered the vocabulary of computer science departments' strategic plans. Innovative courses are being developed that use the arena of experimental computer science to bring industry into tighter contact with academic education. Documentation might include work toward establishing industrial advisory boards and direct consulting with industry to develop specialized courses for industry or new regular courses in the undergraduate curriculum. Service Although faculty in all disciplines have service responsibilities, those in ECSE have two special challenges. The first is that ECSE faculty may be asked to provide service to the university by developing, maintaining, or upgrading software to be used by other members of the university community not in the faculty member's immediate sphere of research interest. In other cases, ECSE faculty are asked to serve on numerous committees to solve computer-related problems faced by the university. Networking experts find themselves on task forces to develop an appropriate networking infrastructure for the campus. Database experts are asked to serve on committees to help solve registration problems, and any ECSE faculty member might be appointed to a committee to select the best system for automating the library. In general, the committee believes that administrations should be discouraged from asking junior ECSE faculty to perform such roles in addition to the demands for committee service placed on all faculty members. However, such service may be appropriate when the overall service load to the university for ECSE faculty (including such computer-related work) is commensurate with that of faculty in other disciplines.

OCR for page 82
Academic Careers for Experimental Computer Scientists and Engineers The second special challenge results from service at the national level. Service includes participation in standards committees, service on advisory boards, invitations to serve on national task forces (including committees of the Computer Science and Telecommunications Board) to solve specific problems in computer science, service to computer societies, and service as reviewers of publications and grant proposals. ECSE has far fewer faculty than older scientific disciplines such as physics or chemistry, and the demands made on senior ECSE faculty for national service are substantial. INSTITUTIONAL CONTRIBUTIONS TO THE ENVIRONMENT In addition to orchestrating a supportive intellectual environment for ECSE and evaluating ECSE faculty in accordance with the standards that characterize the field, departments and universities are directly responsible for certain other aspects of the environment: Start-up funding. As noted earlier, beginning assistant professors are rarely able to secure outside funding in their first year. In other experimental sciences, even new assistant professors are often provided with a start-up package when hired that includes a laboratory, equipment, and other resources. Many universities with ECSE programs do not offer start-up packages at all or, at best, offer incoming faculty (both theoretical and experimental) a single workstation. A start-up package should be offered to beginning assistant professors that is large enough to begin research immediately. At the least, start-up packages for junior ECSE faculty should, when justified by the needs of the research that is planned, be comparable in total dollar value to those offered incoming junior faculty in the more traditional laboratory sciences. Commitment of resources. Departments must understand that high-quality ECSE research with a great impact often demands a substantial commitment of resources. In particular, they must pay careful attention to the following: Equipment upgrades, which for state-of-the-art systems may be necessary as often as yearly; Equipment maintenance; Laboratory space; Technical staff, to keep the computing environment current; Software resources such as computer-aided design (CAD) tools; and Hardware resources such as networking and servers.

OCR for page 82
Academic Careers for Experimental Computer Scientists and Engineers Intellectual property issues. Given the intimate connection between ECSE researchers in academia and industry, departments and universities should promote these interactions. However, one of the most time-consuming aspects of developing these relationships is the resolution of matters related to potential intellectual property arrangements with industrially supported research, or research undertaken jointly with industry. All too often, a faculty researcher lines up an industrial partner, only to see the actual start of research delayed by many months while the university's lawyers negotiate with the company in question. A standard policy or prenegotiated umbrella agreements would go a long way toward facilitating academic-industrial research cooperation.3 Teaching assistant support. As noted earlier, grading and the maintenance of equipment and software place great time demands on ECSE faculty teaching time-intensive, laboratory-based courses. Adequate teaching assistant support for these courses is necessary if the faculty member is to maintain a full portfolio of professional activities. Teaching assignments. Faculty starting a new ECSE research program will need to build teams to carry out the work. Departments can help by providing opportunities for such faculty members to teach advanced seminars in which graduate students can receive needed training in preparation for joining a research project. 3   At universities inexperienced in dealing with licensing and patenting activities, the potentially profitable research activities of ECSE faculty may generate conflict between the faculty and the university administration. As one faculty member put it, Not only are experimentalists subject to undue pressure to get funding, but they are also the subject of university desire to raise funding through commercial ventures. Technology produced by experimentalists is often the subject matter of patenting and licensing activities. This opens a whole new realm of ''back-door university politics" that is unpleasant at best. Universities have yet to learn how to treat and deal with their faculty that produce valuable artifacts.