- Email: [email protected]
Software engineering education: How far we’ve come and how far we have to go
The Journal of Systems and Software 82 (2009) 571–575
Contents lists available at ScienceDirect
The Journal of Systems and Software journal homepage: www.elsevier.com/locate/jss
Software engineering education: How far we’ve come and how far we have to go Nancy R. Mead * Software Engineering Institute, 4500 Fifth Avenue, Pittsburgh, PA 15213, United States
a r t i c l e
i n f o
Article history: Received 2 August 2008 Received in revised form 21 November 2008 Accepted 27 December 2008 Available online 4 January 2009 Keywords: Software engineering education Software engineering history
a b s t r a c t In this paper I trace the history of software engineering education and focus on some of the key players. I highlight what has been accomplished in degree programs and curricula, conferences and working groups, professionalism, certification, and industry–university collaboration. I also look at the challenges that lie ahead—the global reach of education, new delivery mechanisms, new professional efforts, and the need to engage in leadership in software engineering education. What new approaches should be considered? How can we educators maintain our vitality? How can we best nurture new educators and encourage others to join our profession? Ó 2009 Carnegie Mellon University. Published by Elsevier. All rights reserved.
1. Introduction This paper is reminiscent of A Christmas Carol. I feel like I am writing about the ghost of Christmas past, present, and future! In some regards it’s hard to believe what has been accomplished in a relatively short time. On the other hand, it feels like as educators we have barely scratched the surface in terms of what can be done to further software engineering education and to focus on the next generation (otherwise known as the ‘‘net generation”) of software engineers. I’ll start by taking a look at our past accomplishments. Then I’ll take a look at the exciting developments taking place right now. The timeline in Fig. 1 highlights some of those milestones. Finally I’ll get out the crystal ball and try to guess what lies in the future. In the process of doing this, I’ll remember the many contributions made by leaders in the past, take a look at what the current leadership is accomplishing, and guess at the makeup of our leaders in the future. Needless to say, so much has happened that this short paper is far from complete. I am just scratching the surface in this list of accomplishments and have mentioned only a few of the people who have been so influential in the field. 2. The past 2.1. The early days The term software engineering first appeared at the NATO Software Engineering Conference in 1968 (Naur and Randell, 1969; Wikipedia, 2008b). At that time, folks had no idea how controver* Tel.: +1 412 268 5756; fax: +1 412 268 5758. E-mail address: [email protected].
sial the ‘‘engineering” part of this term would be (Dijkstra, 1993). As educators we started to see early publications about structured programming, including the famous ‘‘go-to considered harmful” discussion (Dijkstra, 1968; Wikipedia, 2008a). It’s hard to believe it’s been that long since the moon landing, but it certainly says something about how long software engineering and software engineering education have been around. I was first exposed to software engineering education when I was a practitioner at IBM in the 1970s. (Of course at that time, the term software engineering was still not in common use.) Harlan Mills had convinced the then president of IBM Federal Systems Division, John Jackson, to offer a series of courses entitled Structured Programming Workshop, Structured Design Workshop, and eventually Advanced Design Workshop, fondly known as SPW, SDW, and ADW. SPW and SDW had exams, which determined whether or not you passed the course. SPW emphasized not only structured programming but also formal correctness proofs. Every programmer in the division was to take SPW, which was a two week course preceded by a self-study and exam in logic. Lead programmers were to take SDW, which focused on the use of the state machine model to express abstractions, and the elite would eventually take ADW. ADW focused on issues of concurrency and timeliness, as represented by networks of state machines and associated operations. Eventually these course offerings were supplemented by university short courses. Although it does not seem like much now, taking programmers off the job for two to three weeks and housing them at a central location was a big commitment at the time. The division had around 10–12 thousand staff members then, of which some 2500 or so were programmers. The other factor was that people in the industry did not change jobs very often, so there was an opportunity for the company to reap the benefits of having better educated and trained practitioners. Workshop completions were tracked at
0164-1212/$ - see front matter Ó 2009 Carnegie Mellon University. Published by Elsevier. All rights reserved. doi:10.1016/j.jss.2008.12.038
572
N.R. Mead / The Journal of Systems and Software 82 (2009) 571–575
1968 the term software engineering first appears “go-to considered harmful” discussion published 1970s IBM offers structured programming and design training first MSE programs offered first software engineering books published 1980s Software Engineering Institute founded at Carnegie Mellon first Conference on SE Education takes place MSE model curriculum developed 1990s Forum for the Advancement of SE Education starts SEI CSEE becomes IEEE Computer Society CSEET undergraduate SE degree programs accredited industry-university collaborations grow Texas Board votes to license software engineers distance learning enables global SE education IEEE-CS Software Engineering Code of Ethics established 2000s IEEE-CS adopts SWEBOK ASEET becomes part of CSEET IEEE-CS offers CSDP certification program many universities offer international SE programs SE education tracks introduced in other conferences Fig. 1. A Timeline of Software Engineering Education.
the highest levels of management. The workshops resulted in the Linger–Mills–Witt book on structured programming (Linger et al., 1979), which is still in use. Many of the early pioneers from that era are still working, such as Rick Linger, but others such as Mills and Dijkstra have passed on, having revolutionized the field. In the late 1970s, master’s degree programs in software engineering started to emerge. Ardis was involved in one of the very early programs at Wang Institute of Graduate Studies. Other early programs appeared at Seattle University and Texas Christian University. The workshop series that was so successful in IBM’s Federal Systems Division grew into a corporate-level commitment, and IBM’s Software Engineering Institute was founded in New York City (it was later moved to Westchester County). The corporate model was slightly different, but many of the same courses were offered across the company, and at the corporate-level, in-house instructors were certified to teach them via a formal instructor training program. Fairley wrote one of the first books on software engineering concepts (Fairley, 1984) and went on to become a leader in software engineering education. 2.2. MSE programs, conferences, and workshops During the early 1980s, these educational activities continued, and a study was undertaken that resulted in the current Software Engineering Institute at Carnegie Mellon University. One of the key elements of the original SEI charter was to advance software engineering education. The champion of the SEI at Carnegie Mellon was Angel Jordan, who was the provost. Mary Shaw, at the time a professor in the school of computer science, was heavily involved in the proposal, and Nico Habermann, the dean of the School of Computer Science at Carnegie Mellon, also contributed to the early days of the SEI. Norm Gibbs was hired as Director of Education. Under his leadership, the first Conference on Software Engineering Education (CSEE) took place in 1986/1987. Many of our current leaders were involved in these early activities, both inside and outside the SEI.
Neil Eastman, who was in my management chain at IBM Federal Systems, was involved in development of the Request for Proposal for the SEI and then was on the SEI’s Board of Visitors. At the time of the first CSEE, with Neil’s encouragement, my staff members and I co-authored two papers on software engineering education at IBM and submitted it to one of these early conferences, where they were published as part of the Issues in Software Engineering Education (Hall and Miklos, 1989). Unfortunately, I did not have the opportunity to present the paper, as the conference was too close to my wedding date! As a consequence of these early workshops, and work at the SEI, the initial MSE model curriculum was developed, and an MSE program was started at Carnegie Mellon University in 1990 that continues to this day. Needless to say, the program has undergone changes in content and faculty since those early offerings. The initial faculty included many educators who are still teaching, some who have retired, and sadly, some who have passed on. The latter group includes Norm Gibbs and Jim Tomayko. Jim was responsible for development of the MSE Studio and remained involved in the MSE and the Studio for the rest of his career. He was succeeded by Tony Lattanze, Mel Rosso-Llopart, Gil Taran, and Dave Root. The MSE graduated its first class in 1991. The number of MSE programs grew rapidly, and during these years the SEI developed and distributed instructional materials in the form of videotaped courses, curriculum modules, and educational materials. Many universities used these early materials as a springboard, whereas others developed programs that were uniquely their own. Over time, the various programs came to reflect the specialties and emphases of their universities (Bagert and Mu, 2005). One of the hallmarks of the MSE program is that it was aimed at software practitioners who were already working in the field. As a consequence, the SEI started a continuing education group to focus on the needs of practitioners and managers. This group initially worked on developing a set of courses that were parallel to the MSE courses (Mead and Lawlis, 1991) but later branched out into courses that were unique to practitioners in the field but not nec-
N.R. Mead / The Journal of Systems and Software 82 (2009) 571–575
essarily part of a degree program. Some of these courses were also videotaped and distributed. Train-the-trainer courses were offered so that in-house trainers could deliver the courses to their practitioners. This is the group that I joined in 1990, and later I became Director of SE Education at the SEI, which at the time included workshops, video courses, academic education, and most of the MSE faculty members. In the early 1990s, the CSEE became the catalyst for a newsletter, the Forum for the Advancement of Software Engineering Education (FASE). Over the years it provided informal information to software engineering educators on a periodic basis. It continues to provide announcements and articles to its audience, which is worldwide. The editors included Keith Pierce, Don Bagert, and the current editors, Barrie Thompson and Helen Edwards. In the early days of the internet, producing and distributing an electronic newsletter was quite a challenge, as Don can attest to. The CSEE conference series also broadened its focus to include training and as of 1997 became known as the Conference on Software Engineering Education and Training (CSEET). This was not so much a change of emphasis as recognition of the fact that education and training have always gone hand in hand. The conference also transitioned from an SEI conference to an IEEE conference. This was no mean feat, as it involved replacing the Events staff at the SEI with volunteers and securing IEEE sponsorship. I was the first Steering Committee chair, a glutton for punishment. Subsequent chairs were Don Bagert, and now Tim Lethbridge. 2.3. Industry–university collaboration and the WGSEET In 1995, a working group for software engineering education and training (WGSEET) was formed under my leadership. This was an ad hoc group whose goal was to advance software engineering education. The WGSEET provided a forum for advancing the field and for tackling controversial issues in a friendly environment. It continued for a number of years as a standalone group, publishing many useful papers and reports. Its topics included curriculum development, industry–university collaboration, and professionalism. Eventually it was disbanded in favor of work under the umbrella of IEEE-CS. The WGSEET provided an opportunity to support software engineering curriculum development and publication. In fact, the work of the WGSEET preceded software engineering curriculum development efforts by the professional societies, led by professionals such as Rich LeBlanc and Gerry Engel. Later on members of the group applied for a grant to develop a repository for educational materials (SWENET). About the same time, undergraduate degree programs in software engineering education started to emerge, and in the US a merger of the accrediting bodies, CSAB and ABET, made it possible for them to be accredited. New curricula were developed to support a variety of computing degree programs (Joint Taskforce for Computing Curricula, 2004). Universities such as Rochester Institute of Technology, under the leadership of Mike Lutz, were pioneers in offering the first undergraduate degrees in software engineering in the US (Lutz, 1996). Other countries such as Australia found it easier to offer undergraduate degrees in software engineering. Their accreditation processes examined engineering programs as a whole, and so they did not have to wrestle as much with the question of whether software engineering was a legitimate engineering discipline. In the EU many of the software engineering offerings were in departments that taught Computing, so once again the term ‘‘engineering” was not problematic for these universities. It became clear that part of the agenda was to collaborate with industry as well as influence it. The early MSE programs were
573
among the roots of work in industry–university collaboration. Many universities established industry advisory boards so that their degree programs could remain relevant to their industry partners. These collaborations took many forms—joint research, regular work sessions or meetings, and the like. This provided an opportunity to study these collaborations and try to figure out what made them tick (Beckman et al., 2000; Irons and Alexander, 2004; Beckman et al., 1997; Ellis et al., 2001; Mead, 1996). Working Group members collaborated to publish a number of papers and reports on industry–university collaboration. These included not only efforts at US universities and their industry counterparts, but we found significant industry–university collaboration taking place in Madrid, Spain (Ellis et al., 2001). The WGSEET met twice a year—once in conjunction with CSEET, and one other time, usually in conjunction with another conference such as Frontiers in Education. However, industry–university collaboration was never a controversial issue. It took the discussion of undergraduate degree programs in software engineering and the idea of licensing software professionals to generate true controversy! A joint committee was formed by ACM and IEEE Computer Society in 1993 to promote software engineering as a profession. Dennis Frailey, who has been involved with the CSEET from the beginning, was co-chair of the initial IEEE/ACM Steering Committee to Establish Software Engineering as a Profession (1994–1998) (and then vice-chair of the Software Engineering Coordinating Committee from 1999–2001.) The SWECC supported development of the SWEBOK (IEEE-CS, 2008b). The SWEBOK was a multiyear effort, with Robert Dupuis and Pierre Bourque serving as primary editors. There was an extensive process of multiple reviews, and an industrial advisory board. Many professionals contributed to it. Once the SWEBOK was completed, some professionals saw it as the basis for certification programs, but others feared that it would become the basis for licensing. In that same timeframe, Don Gotterbarn led an effort to establish the Software Engineering Code of Ethics and Professional Practice (IEEE-CS, 1999) and an associated curriculum. His dedication to this effort in the face of controversy and his perseverance are noteworthy. Licensing generated serious controversy among thoughtful people that continues to this day. Papers about licensing generated a level of rhetoric that had not been seen before (Mead, 1997b, 1998). More scholarly articles received fewer slings and arrows, but were nevertheless controversial (Mead and Turner, 1998). Don Bagert was the subject of much discussion when he became the first licensed software engineer in the state of Texas. Licensing has not become extensive in the US, but it successfully faced legal challenges in Canada. The discussion continues, and the future for licensing is unclear (Boykin, 2007). IEEE-CS started to offer a certification program that was initially based on the SWEBOK and has evolved to include several levels of certification (IEEE-CS, 2008a). The exams were professionally developed and tested extensively. This program is less controversial than licensing, but participation in it has also been fairly light. The term Certified Software Development Professional (CSDP) avoids the use of the term engineer. 2.4. Delivery mechanisms Distance learning became a popular concept (Mead, 1997a), as well as the establishment of branch locations among universities. Distance learning initially relied on the use of videotaped material, along with instructor notes, assignments, exams, etc. Later on there was a focus on satellite delivery of live lectures, and more recently web-based course offerings, supplemented by online chat sessions have become popular. Initially there were videotapes of actual
574
N.R. Mead / The Journal of Systems and Software 82 (2009) 571–575
lectures, which were distributed after the fact. A good bit of interaction was required in order to do this in a timely manner, and the lectures could be fairly dry from the point of view of the students viewing the tapes. One attempt to improve quality resulted in ‘‘direct address” delivery, where the instructor addressed a monitor in a studio with no students, similar to what TV broadcasters do. However, not all instructors were comfortable with talking to a camera without any live student interaction or feedback. Satellite delivery was not a whole lot better, as there were often a combination of local and remote students in the class, and the remote students were often reluctant to jump in and ask questions. Web-based delivery seems to have overcome some of these early difficulties, as it is a multimedia experience and does not rely as extensively on video material. Also, the exams and assignments can be handled asynchronously. Globalization became a buzzword as universities tried to outdo each other in the number of far-flung locations that they supported. Going global is not as easy as it might first appear. The university must find a location that will support a critical mass of students and then must find suitable classroom facilities. Faculty must either be hired or brought in from their home location on a temporary basis. There may be significant cultural differences in the global location(s) as compared to the university’s main campus. Often students will complete part of their degree program at their home location and part of their degree program at the university’s main campus. This is an enriching experience for the global students, the local students, and the faculty. Although it is not easy to do, there seem to be significant benefits to be gained from globalization.
international efforts, either on-site or as a form of distance education. Distance delivery via a variety of mechanisms is routine. Creative programs allow working students almost anywhere to complete many, and in some cases all, of their degree requirements at a distance. The number of conferences and conference tracks also has grown. The ACM Special Interest Group on Computer Science Education has long sponsored a conference series (SIGCSE, 2008) that has contained software engineering content from time to time. The ITiCSE (Innovation and Technology in Computer Science Education) Conference (ITiCSE, 2008) has also contained software engineering content. The series of Conferences on the Teaching of Computing (CTC) in Ireland, under the leadership of Michael Ryan and Micheal O’hEigeartaigh, had a considerable amount of software engineering content. More recently, ICSE (International Conference on Software Engineering) has included a software engineering education track (ICSE, 2008). In the year 2000, Mary Shaw presented a discussion of the future of software engineering education as part of a special track at ICSE (Shaw, 2000). Frontiers in Education (FIE) has always included software engineering content.
3. The present
A recent effort under the leadership of Art Pyster and Rich Turner at Stevens Institute of Technology, Graduate Software Engineering Reference Curriculum (GSwERC), has resulted in the Integrated Software and Systems Engineering Curriculum. Many of the folks who have been involved in development of software engineering curricula are involved in this effort, which is a welcome addition. This project aims at a master’s curriculum that includes software and systems engineering. At the 2007 CSEET conference, Dave Parnas said that he has come to think that maybe there are three types of degree programs: computer science for researchers, software engineering for engineers developing mission–critical systems, and software professional degrees for practitioners developing non-critical systems (Parnas, 2007). The trend toward globalization is here to stay. We will continue to see creative evolution of delivery mechanisms. We can also expect to see multicultural teams and offshoring. Of course the meaning of the term offshoring depends on where you are located. Offerings such as the ASEET will help to develop a new generation of software engineering educators. Conferences, working groups, and committees will evolve. Electronic publishing is becoming much more common, as are video conferences and webcasts. I get webcast announcements all the time, and I suspect that most of you do too. Web-based collaboration mechanisms have also become popular. Professionalism continues to be a hot topic, and it’s still unpredictable whether licensing or certification will become a major trend in the future. Related to that are discussions about whether litigation over software quality, safety, and security will become more common. New specializations will emerge. For example, software security and software assurance have become big concerns for many of us, and these as well as other new directions will result in change.
The CSEET has evolved to include the ASEET, the Academy for Software Engineering Educators and Trainers. This is a wonderful addition to the conference and supports the idea of mentoring as a way of growing the skills that are needed for professional software engineering education. It is truly an exciting development. Dan Port was involved in the initial proposal for the ASEET, and Tom Horton was the chair of ASEET 2008. IN 2008 there were around 30 participants, and the speakers included Barry Boehm as keynote, and Vic Basili, Jared Richardson, and Dieter Rombach as instructors. One bold move in the early days of CSEET was to move the conference out of Pittsburgh. After many years in Pittsburgh, our first venture to another location was the 1992 conference in San Diego. This particular conference, the 6th CSEE, was chaired by Carol Sledge, while still under the auspices of the SEI, and boasted an attendance of over 200. Another bold move was to start holding the conference in international locations, such as Spain, Canada, and Ireland. Many of the early pioneers on the CSEET Steering Committee have been succeeded by dynamic young educators. This is all to the good. New degree programs are springing up all the time. PhD programs in software engineering exist at a number of universities, including Carnegie Mellon, US Naval Postgraduate School (NPS), and Lero (Software Engineering for Ireland), among others. Graduates of these programs will be well positioned to enter into software engineering research and software engineering education. At Carnegie Mellon, there are now a number of international locations that are offering unique degree programs, including Australia, Greece, Japan, Portugal, and Qatar (Carnegie Mellon, 2008). New degree programs, such as the MS in Information Technology – Software Engineering Management, the MSIT-SEM, continue to be defined. The MSIT-SEM has been influenced by the SEI’s work in the process area. Many other universities have undertaken similar
4. The future Although it is interesting to reflect on the past, we do not want to get stuck there. Nor do we want to become complacent in the present. Change is happening all around us, and we need to adjust to and motivate change. Here are some possible indicators of future trends:
N.R. Mead / The Journal of Systems and Software 82 (2009) 571–575
5. Conclusion So what does all this mean to us? It means that we will have continued opportunities to provide leadership as software engineering education evolves and matures. Here are some of the challenges that we need to undertake as educators: We will be called upon to mentor and support new faculty who are challenged to educate the ‘‘net generation” of software engineers. Software engineering PhD programs will be one key to producing new faculty in software engineering. Offerings such as the ASEET will help to develop a new generation of software engineering educators and will serve to support mentoring activities. We need to continue to encourage industry–university collaboration. It will result in joint research and education projects and will keep our educational offerings grounded in practice. Although we have observed many such successful collaborations, they have not become as commonplace as we would have hoped. We would urge university departments and industry organizations with an interest in software engineering to seek out each other and start this important dialogue. Many of us have been challenged by the fact that educational research is not recognized as ‘‘legitimate” technical research. It somehow does not ‘‘count” when it comes to publications, conferences, tenure, and so on, except at those universities with a strong commitment to education as a legitimate area of research in its own right. We need to continue to push for recognition of educational research as a legitimate field of research. We need to further the profession. We need to recognize the need to be change agents and to motivate the next generation of software engineers and software engineering educators. Software engineering specializations within a broad variety of degree programs will help to reach this education goal. Professional certifications such as the IEEE Certified Software Development Professional (CSDP) as well as professional society memberships may help in this area. We are seeing specializations in related areas, such as information security, software security, and software assurance, as well as certifications in these areas by SANS and others. These are clearly aimed at security of our software and networks. We also need to embrace the IEEE Code of Ethics (IEEE, 2006) and use it to guide our work. These are big challenges, but I believe that if we work together as educators we can be successful in meeting the challenge. Our challenge is to exercise leadership by making things happen, not just by doing our job. It would not be easy, but it will be fun!
References Bagert, D.J., Mu, X., 2005. Current state of software engineering master’s degree programs in the United States. In: Proc. 35th Annu. Conf. Frontiers in Education Conference, IEEE Computer Society Press, New York, F1G-1–F1G-6.. Beckman, K., Coulter, N., Khajenoori, S., Mead, N.R., 1997. Collaborations: Closing the industry–academia gap. IEEE Software 14 (6), 49–57. Beckman, K., Lawrence, J., Mead, N., O’Mary, G., Parish, C., Unpingco, P., Walker, H., 2000. Industry/university collaborations: different perspectives heighten mutual opportunities. CrossTalk 13 (3), 10–15.
575
Boykin, D., 2007. Is it time to license software engineers? PE Magazine.. Carnegie Mellon University. 2008. Carnegie Mellon: The Global University. . Dijkstra, E.W., 1968. Letters to the editor: go to statement considered harmful. Communications of the ACM 11 (3), 147–148. doi:10.1145/362929.362947. Dijkstra, E.W., 1993. There is still a war going on, E.W. Dijkstra Archive. . Ellis, H.J.C., Mead, N.R., Moreno, A., MacNeil, P., 2001. Can industry and academia collaborate to meet the need for software engineers? Cutter IT Journal 14 (6), 32–39. Fairley, R.E., 1984. Software Engineering Concepts. McGraw Hill, New York. Hall, N., Miklos, J., 1989. Formal education within the software life cycle. In: Fairley, R., Freeman, P. (Eds.), Issues in Software Eng. Ed. Springer-Verlag, New York. ICSE. 2008. Proceedings of the 30th International Conference on Software Engineering 2008, Leipzig, Germany, May 10–18. ACM Publications. IEEE Computer Society. 1999. Software Engineering Code of Ethics and Professional Practice. . IEEE Computer Society. 2008a. CSDP: Certified Software Development Professional. . IEEE Computer Society. 2008b. SWEBOK Overview, SE Online. . Irons, A., Alexander, S. (Eds.), 2004. Industrial input to the computing curriculum. Effective Learning and Teaching in Computing. Routledge Falmer, New York, pp. 123–135. ITiCSE. 2008. In: Amillo, J., Laxer, C., Menasalvas Ruiz, E., Young, A. (Eds.): Proc. 13th Annu. SIGCSE Conf. Innovation and Technology in Computer Science Education, ITiCSE 2008, Madrid, Spain, June 30–July 2, 2008. ACM 2008, ISBN 978-160558-078-4. Joint Taskforce for Computing Curricula. 2004. Software Engineering 2004: Curricular Guidelines for Undergraduate Programs in Software Engineering, ACM/IEEE, New York. Linger, R.C., Mills, H.D., Witt, B.I., 1979. Structured Programming: Theory and Practice. Addison-Wesley, Reading, Massachusetts. Lutz, M., 1996. RIT offers nation’s first undergraduate degree in software engineering, Forum for Advancing Software Eng. Ed. vol. 6, p. 14. . Mead, N.R., 1996. Industry–university partnerships: The wave of the future? In: Proc. 4th Annu. Conf. Teaching of Computing, Centre for Teaching Computing, Dublin City University, Dublin, Ireland. Mead, N.R., 1997a. Distance learning: Lessons learned. In: 5th Annu. Conf. Teaching of Computing, Centre for Teaching Computing, Dublin City University, Dublin. Mead, N.R., 1997b. Issues in licensing and certification of software engineers. In: Proc. 10th Conf. Software Eng. Ed. and Training. IEEE Computer Society Press, New York, pp. 150–160. Mead, N.R., 1998. Are we going to fish or cut bait? Licensing and certification of software professionals. Cutter IT J. 11 (5), 4–8. Mead, N.R., Lawlis, P., 1991. Software engineering: Graduate-level courses for AFIT professional continuing education. In: Proc. 5th SEI Conf. Software Eng. Ed. Springer-Verlag, New York. Mead, N.R., Turner, A.J., 1998. Current accreditation, certification, and licensure activities related to software engineering. Annals of Software Eng. 6, 167–180. Naur, P., Randell, B. (Eds.), 1969. Software Engineering: Report on a Conference Sponsored by the NATO Science Committee, Garmisch, Germany, October 7–11, 1968, Brussels, Scientific Affairs Division, NATO. Parnas, D., 2007. Resolving dilemmas in software engineering education. In: Proc. 20th Conf. Software Eng. Ed. and Training, IEEE Computer Society Press, New York, xix. doi.ieeecomputersociety.org/10.1109/CSEET.2007.43. Shaw, M., 2000. Software engineering education: a roadmap. ICSE – Future of SE Track 2000, pp. 371–380. . SIGCSE, 2008. In: Proceedings of the 39th SIGCSE Technical Symposium on Computer Science Education 2008, Portland, OR, USA, March 12–15, 2008, ACM. Wikipedia, 2008a. Edsger W. Dijkstra. . Wikipedia, 2008b. Software Engineering. . Nancy R. Mead is a senior member of the technical staff at the Software Engineering Institute and a faculty member in the Master of Software Engineering and Master of Information Systems Management programs at Carnegie Mellon University. She is currently involved in the study of secure systems engineering and the development of professional infrastructure for software engineers. Mead received her PhD in mathematics from the Polytechnic Institute of New York and a BA and an MS in mathematics from New York University. She is a member of the ACM and a Fellow of IEEE and the IEEE Computer Society.
Contents lists available at ScienceDirect
The Journal of Systems and Software journal homepage: www.elsevier.com/locate/jss
Software engineering education: How far we’ve come and how far we have to go Nancy R. Mead * Software Engineering Institute, 4500 Fifth Avenue, Pittsburgh, PA 15213, United States
a r t i c l e
i n f o
Article history: Received 2 August 2008 Received in revised form 21 November 2008 Accepted 27 December 2008 Available online 4 January 2009 Keywords: Software engineering education Software engineering history
a b s t r a c t In this paper I trace the history of software engineering education and focus on some of the key players. I highlight what has been accomplished in degree programs and curricula, conferences and working groups, professionalism, certification, and industry–university collaboration. I also look at the challenges that lie ahead—the global reach of education, new delivery mechanisms, new professional efforts, and the need to engage in leadership in software engineering education. What new approaches should be considered? How can we educators maintain our vitality? How can we best nurture new educators and encourage others to join our profession? Ó 2009 Carnegie Mellon University. Published by Elsevier. All rights reserved.
1. Introduction This paper is reminiscent of A Christmas Carol. I feel like I am writing about the ghost of Christmas past, present, and future! In some regards it’s hard to believe what has been accomplished in a relatively short time. On the other hand, it feels like as educators we have barely scratched the surface in terms of what can be done to further software engineering education and to focus on the next generation (otherwise known as the ‘‘net generation”) of software engineers. I’ll start by taking a look at our past accomplishments. Then I’ll take a look at the exciting developments taking place right now. The timeline in Fig. 1 highlights some of those milestones. Finally I’ll get out the crystal ball and try to guess what lies in the future. In the process of doing this, I’ll remember the many contributions made by leaders in the past, take a look at what the current leadership is accomplishing, and guess at the makeup of our leaders in the future. Needless to say, so much has happened that this short paper is far from complete. I am just scratching the surface in this list of accomplishments and have mentioned only a few of the people who have been so influential in the field. 2. The past 2.1. The early days The term software engineering first appeared at the NATO Software Engineering Conference in 1968 (Naur and Randell, 1969; Wikipedia, 2008b). At that time, folks had no idea how controver* Tel.: +1 412 268 5756; fax: +1 412 268 5758. E-mail address: [email protected].
sial the ‘‘engineering” part of this term would be (Dijkstra, 1993). As educators we started to see early publications about structured programming, including the famous ‘‘go-to considered harmful” discussion (Dijkstra, 1968; Wikipedia, 2008a). It’s hard to believe it’s been that long since the moon landing, but it certainly says something about how long software engineering and software engineering education have been around. I was first exposed to software engineering education when I was a practitioner at IBM in the 1970s. (Of course at that time, the term software engineering was still not in common use.) Harlan Mills had convinced the then president of IBM Federal Systems Division, John Jackson, to offer a series of courses entitled Structured Programming Workshop, Structured Design Workshop, and eventually Advanced Design Workshop, fondly known as SPW, SDW, and ADW. SPW and SDW had exams, which determined whether or not you passed the course. SPW emphasized not only structured programming but also formal correctness proofs. Every programmer in the division was to take SPW, which was a two week course preceded by a self-study and exam in logic. Lead programmers were to take SDW, which focused on the use of the state machine model to express abstractions, and the elite would eventually take ADW. ADW focused on issues of concurrency and timeliness, as represented by networks of state machines and associated operations. Eventually these course offerings were supplemented by university short courses. Although it does not seem like much now, taking programmers off the job for two to three weeks and housing them at a central location was a big commitment at the time. The division had around 10–12 thousand staff members then, of which some 2500 or so were programmers. The other factor was that people in the industry did not change jobs very often, so there was an opportunity for the company to reap the benefits of having better educated and trained practitioners. Workshop completions were tracked at
0164-1212/$ - see front matter Ó 2009 Carnegie Mellon University. Published by Elsevier. All rights reserved. doi:10.1016/j.jss.2008.12.038
572
N.R. Mead / The Journal of Systems and Software 82 (2009) 571–575
1968 the term software engineering first appears “go-to considered harmful” discussion published 1970s IBM offers structured programming and design training first MSE programs offered first software engineering books published 1980s Software Engineering Institute founded at Carnegie Mellon first Conference on SE Education takes place MSE model curriculum developed 1990s Forum for the Advancement of SE Education starts SEI CSEE becomes IEEE Computer Society CSEET undergraduate SE degree programs accredited industry-university collaborations grow Texas Board votes to license software engineers distance learning enables global SE education IEEE-CS Software Engineering Code of Ethics established 2000s IEEE-CS adopts SWEBOK ASEET becomes part of CSEET IEEE-CS offers CSDP certification program many universities offer international SE programs SE education tracks introduced in other conferences Fig. 1. A Timeline of Software Engineering Education.
the highest levels of management. The workshops resulted in the Linger–Mills–Witt book on structured programming (Linger et al., 1979), which is still in use. Many of the early pioneers from that era are still working, such as Rick Linger, but others such as Mills and Dijkstra have passed on, having revolutionized the field. In the late 1970s, master’s degree programs in software engineering started to emerge. Ardis was involved in one of the very early programs at Wang Institute of Graduate Studies. Other early programs appeared at Seattle University and Texas Christian University. The workshop series that was so successful in IBM’s Federal Systems Division grew into a corporate-level commitment, and IBM’s Software Engineering Institute was founded in New York City (it was later moved to Westchester County). The corporate model was slightly different, but many of the same courses were offered across the company, and at the corporate-level, in-house instructors were certified to teach them via a formal instructor training program. Fairley wrote one of the first books on software engineering concepts (Fairley, 1984) and went on to become a leader in software engineering education. 2.2. MSE programs, conferences, and workshops During the early 1980s, these educational activities continued, and a study was undertaken that resulted in the current Software Engineering Institute at Carnegie Mellon University. One of the key elements of the original SEI charter was to advance software engineering education. The champion of the SEI at Carnegie Mellon was Angel Jordan, who was the provost. Mary Shaw, at the time a professor in the school of computer science, was heavily involved in the proposal, and Nico Habermann, the dean of the School of Computer Science at Carnegie Mellon, also contributed to the early days of the SEI. Norm Gibbs was hired as Director of Education. Under his leadership, the first Conference on Software Engineering Education (CSEE) took place in 1986/1987. Many of our current leaders were involved in these early activities, both inside and outside the SEI.
Neil Eastman, who was in my management chain at IBM Federal Systems, was involved in development of the Request for Proposal for the SEI and then was on the SEI’s Board of Visitors. At the time of the first CSEE, with Neil’s encouragement, my staff members and I co-authored two papers on software engineering education at IBM and submitted it to one of these early conferences, where they were published as part of the Issues in Software Engineering Education (Hall and Miklos, 1989). Unfortunately, I did not have the opportunity to present the paper, as the conference was too close to my wedding date! As a consequence of these early workshops, and work at the SEI, the initial MSE model curriculum was developed, and an MSE program was started at Carnegie Mellon University in 1990 that continues to this day. Needless to say, the program has undergone changes in content and faculty since those early offerings. The initial faculty included many educators who are still teaching, some who have retired, and sadly, some who have passed on. The latter group includes Norm Gibbs and Jim Tomayko. Jim was responsible for development of the MSE Studio and remained involved in the MSE and the Studio for the rest of his career. He was succeeded by Tony Lattanze, Mel Rosso-Llopart, Gil Taran, and Dave Root. The MSE graduated its first class in 1991. The number of MSE programs grew rapidly, and during these years the SEI developed and distributed instructional materials in the form of videotaped courses, curriculum modules, and educational materials. Many universities used these early materials as a springboard, whereas others developed programs that were uniquely their own. Over time, the various programs came to reflect the specialties and emphases of their universities (Bagert and Mu, 2005). One of the hallmarks of the MSE program is that it was aimed at software practitioners who were already working in the field. As a consequence, the SEI started a continuing education group to focus on the needs of practitioners and managers. This group initially worked on developing a set of courses that were parallel to the MSE courses (Mead and Lawlis, 1991) but later branched out into courses that were unique to practitioners in the field but not nec-
N.R. Mead / The Journal of Systems and Software 82 (2009) 571–575
essarily part of a degree program. Some of these courses were also videotaped and distributed. Train-the-trainer courses were offered so that in-house trainers could deliver the courses to their practitioners. This is the group that I joined in 1990, and later I became Director of SE Education at the SEI, which at the time included workshops, video courses, academic education, and most of the MSE faculty members. In the early 1990s, the CSEE became the catalyst for a newsletter, the Forum for the Advancement of Software Engineering Education (FASE). Over the years it provided informal information to software engineering educators on a periodic basis. It continues to provide announcements and articles to its audience, which is worldwide. The editors included Keith Pierce, Don Bagert, and the current editors, Barrie Thompson and Helen Edwards. In the early days of the internet, producing and distributing an electronic newsletter was quite a challenge, as Don can attest to. The CSEE conference series also broadened its focus to include training and as of 1997 became known as the Conference on Software Engineering Education and Training (CSEET). This was not so much a change of emphasis as recognition of the fact that education and training have always gone hand in hand. The conference also transitioned from an SEI conference to an IEEE conference. This was no mean feat, as it involved replacing the Events staff at the SEI with volunteers and securing IEEE sponsorship. I was the first Steering Committee chair, a glutton for punishment. Subsequent chairs were Don Bagert, and now Tim Lethbridge. 2.3. Industry–university collaboration and the WGSEET In 1995, a working group for software engineering education and training (WGSEET) was formed under my leadership. This was an ad hoc group whose goal was to advance software engineering education. The WGSEET provided a forum for advancing the field and for tackling controversial issues in a friendly environment. It continued for a number of years as a standalone group, publishing many useful papers and reports. Its topics included curriculum development, industry–university collaboration, and professionalism. Eventually it was disbanded in favor of work under the umbrella of IEEE-CS. The WGSEET provided an opportunity to support software engineering curriculum development and publication. In fact, the work of the WGSEET preceded software engineering curriculum development efforts by the professional societies, led by professionals such as Rich LeBlanc and Gerry Engel. Later on members of the group applied for a grant to develop a repository for educational materials (SWENET). About the same time, undergraduate degree programs in software engineering education started to emerge, and in the US a merger of the accrediting bodies, CSAB and ABET, made it possible for them to be accredited. New curricula were developed to support a variety of computing degree programs (Joint Taskforce for Computing Curricula, 2004). Universities such as Rochester Institute of Technology, under the leadership of Mike Lutz, were pioneers in offering the first undergraduate degrees in software engineering in the US (Lutz, 1996). Other countries such as Australia found it easier to offer undergraduate degrees in software engineering. Their accreditation processes examined engineering programs as a whole, and so they did not have to wrestle as much with the question of whether software engineering was a legitimate engineering discipline. In the EU many of the software engineering offerings were in departments that taught Computing, so once again the term ‘‘engineering” was not problematic for these universities. It became clear that part of the agenda was to collaborate with industry as well as influence it. The early MSE programs were
573
among the roots of work in industry–university collaboration. Many universities established industry advisory boards so that their degree programs could remain relevant to their industry partners. These collaborations took many forms—joint research, regular work sessions or meetings, and the like. This provided an opportunity to study these collaborations and try to figure out what made them tick (Beckman et al., 2000; Irons and Alexander, 2004; Beckman et al., 1997; Ellis et al., 2001; Mead, 1996). Working Group members collaborated to publish a number of papers and reports on industry–university collaboration. These included not only efforts at US universities and their industry counterparts, but we found significant industry–university collaboration taking place in Madrid, Spain (Ellis et al., 2001). The WGSEET met twice a year—once in conjunction with CSEET, and one other time, usually in conjunction with another conference such as Frontiers in Education. However, industry–university collaboration was never a controversial issue. It took the discussion of undergraduate degree programs in software engineering and the idea of licensing software professionals to generate true controversy! A joint committee was formed by ACM and IEEE Computer Society in 1993 to promote software engineering as a profession. Dennis Frailey, who has been involved with the CSEET from the beginning, was co-chair of the initial IEEE/ACM Steering Committee to Establish Software Engineering as a Profession (1994–1998) (and then vice-chair of the Software Engineering Coordinating Committee from 1999–2001.) The SWECC supported development of the SWEBOK (IEEE-CS, 2008b). The SWEBOK was a multiyear effort, with Robert Dupuis and Pierre Bourque serving as primary editors. There was an extensive process of multiple reviews, and an industrial advisory board. Many professionals contributed to it. Once the SWEBOK was completed, some professionals saw it as the basis for certification programs, but others feared that it would become the basis for licensing. In that same timeframe, Don Gotterbarn led an effort to establish the Software Engineering Code of Ethics and Professional Practice (IEEE-CS, 1999) and an associated curriculum. His dedication to this effort in the face of controversy and his perseverance are noteworthy. Licensing generated serious controversy among thoughtful people that continues to this day. Papers about licensing generated a level of rhetoric that had not been seen before (Mead, 1997b, 1998). More scholarly articles received fewer slings and arrows, but were nevertheless controversial (Mead and Turner, 1998). Don Bagert was the subject of much discussion when he became the first licensed software engineer in the state of Texas. Licensing has not become extensive in the US, but it successfully faced legal challenges in Canada. The discussion continues, and the future for licensing is unclear (Boykin, 2007). IEEE-CS started to offer a certification program that was initially based on the SWEBOK and has evolved to include several levels of certification (IEEE-CS, 2008a). The exams were professionally developed and tested extensively. This program is less controversial than licensing, but participation in it has also been fairly light. The term Certified Software Development Professional (CSDP) avoids the use of the term engineer. 2.4. Delivery mechanisms Distance learning became a popular concept (Mead, 1997a), as well as the establishment of branch locations among universities. Distance learning initially relied on the use of videotaped material, along with instructor notes, assignments, exams, etc. Later on there was a focus on satellite delivery of live lectures, and more recently web-based course offerings, supplemented by online chat sessions have become popular. Initially there were videotapes of actual
574
N.R. Mead / The Journal of Systems and Software 82 (2009) 571–575
lectures, which were distributed after the fact. A good bit of interaction was required in order to do this in a timely manner, and the lectures could be fairly dry from the point of view of the students viewing the tapes. One attempt to improve quality resulted in ‘‘direct address” delivery, where the instructor addressed a monitor in a studio with no students, similar to what TV broadcasters do. However, not all instructors were comfortable with talking to a camera without any live student interaction or feedback. Satellite delivery was not a whole lot better, as there were often a combination of local and remote students in the class, and the remote students were often reluctant to jump in and ask questions. Web-based delivery seems to have overcome some of these early difficulties, as it is a multimedia experience and does not rely as extensively on video material. Also, the exams and assignments can be handled asynchronously. Globalization became a buzzword as universities tried to outdo each other in the number of far-flung locations that they supported. Going global is not as easy as it might first appear. The university must find a location that will support a critical mass of students and then must find suitable classroom facilities. Faculty must either be hired or brought in from their home location on a temporary basis. There may be significant cultural differences in the global location(s) as compared to the university’s main campus. Often students will complete part of their degree program at their home location and part of their degree program at the university’s main campus. This is an enriching experience for the global students, the local students, and the faculty. Although it is not easy to do, there seem to be significant benefits to be gained from globalization.
international efforts, either on-site or as a form of distance education. Distance delivery via a variety of mechanisms is routine. Creative programs allow working students almost anywhere to complete many, and in some cases all, of their degree requirements at a distance. The number of conferences and conference tracks also has grown. The ACM Special Interest Group on Computer Science Education has long sponsored a conference series (SIGCSE, 2008) that has contained software engineering content from time to time. The ITiCSE (Innovation and Technology in Computer Science Education) Conference (ITiCSE, 2008) has also contained software engineering content. The series of Conferences on the Teaching of Computing (CTC) in Ireland, under the leadership of Michael Ryan and Micheal O’hEigeartaigh, had a considerable amount of software engineering content. More recently, ICSE (International Conference on Software Engineering) has included a software engineering education track (ICSE, 2008). In the year 2000, Mary Shaw presented a discussion of the future of software engineering education as part of a special track at ICSE (Shaw, 2000). Frontiers in Education (FIE) has always included software engineering content.
3. The present
A recent effort under the leadership of Art Pyster and Rich Turner at Stevens Institute of Technology, Graduate Software Engineering Reference Curriculum (GSwERC), has resulted in the Integrated Software and Systems Engineering Curriculum. Many of the folks who have been involved in development of software engineering curricula are involved in this effort, which is a welcome addition. This project aims at a master’s curriculum that includes software and systems engineering. At the 2007 CSEET conference, Dave Parnas said that he has come to think that maybe there are three types of degree programs: computer science for researchers, software engineering for engineers developing mission–critical systems, and software professional degrees for practitioners developing non-critical systems (Parnas, 2007). The trend toward globalization is here to stay. We will continue to see creative evolution of delivery mechanisms. We can also expect to see multicultural teams and offshoring. Of course the meaning of the term offshoring depends on where you are located. Offerings such as the ASEET will help to develop a new generation of software engineering educators. Conferences, working groups, and committees will evolve. Electronic publishing is becoming much more common, as are video conferences and webcasts. I get webcast announcements all the time, and I suspect that most of you do too. Web-based collaboration mechanisms have also become popular. Professionalism continues to be a hot topic, and it’s still unpredictable whether licensing or certification will become a major trend in the future. Related to that are discussions about whether litigation over software quality, safety, and security will become more common. New specializations will emerge. For example, software security and software assurance have become big concerns for many of us, and these as well as other new directions will result in change.
The CSEET has evolved to include the ASEET, the Academy for Software Engineering Educators and Trainers. This is a wonderful addition to the conference and supports the idea of mentoring as a way of growing the skills that are needed for professional software engineering education. It is truly an exciting development. Dan Port was involved in the initial proposal for the ASEET, and Tom Horton was the chair of ASEET 2008. IN 2008 there were around 30 participants, and the speakers included Barry Boehm as keynote, and Vic Basili, Jared Richardson, and Dieter Rombach as instructors. One bold move in the early days of CSEET was to move the conference out of Pittsburgh. After many years in Pittsburgh, our first venture to another location was the 1992 conference in San Diego. This particular conference, the 6th CSEE, was chaired by Carol Sledge, while still under the auspices of the SEI, and boasted an attendance of over 200. Another bold move was to start holding the conference in international locations, such as Spain, Canada, and Ireland. Many of the early pioneers on the CSEET Steering Committee have been succeeded by dynamic young educators. This is all to the good. New degree programs are springing up all the time. PhD programs in software engineering exist at a number of universities, including Carnegie Mellon, US Naval Postgraduate School (NPS), and Lero (Software Engineering for Ireland), among others. Graduates of these programs will be well positioned to enter into software engineering research and software engineering education. At Carnegie Mellon, there are now a number of international locations that are offering unique degree programs, including Australia, Greece, Japan, Portugal, and Qatar (Carnegie Mellon, 2008). New degree programs, such as the MS in Information Technology – Software Engineering Management, the MSIT-SEM, continue to be defined. The MSIT-SEM has been influenced by the SEI’s work in the process area. Many other universities have undertaken similar
4. The future Although it is interesting to reflect on the past, we do not want to get stuck there. Nor do we want to become complacent in the present. Change is happening all around us, and we need to adjust to and motivate change. Here are some possible indicators of future trends:
N.R. Mead / The Journal of Systems and Software 82 (2009) 571–575
5. Conclusion So what does all this mean to us? It means that we will have continued opportunities to provide leadership as software engineering education evolves and matures. Here are some of the challenges that we need to undertake as educators: We will be called upon to mentor and support new faculty who are challenged to educate the ‘‘net generation” of software engineers. Software engineering PhD programs will be one key to producing new faculty in software engineering. Offerings such as the ASEET will help to develop a new generation of software engineering educators and will serve to support mentoring activities. We need to continue to encourage industry–university collaboration. It will result in joint research and education projects and will keep our educational offerings grounded in practice. Although we have observed many such successful collaborations, they have not become as commonplace as we would have hoped. We would urge university departments and industry organizations with an interest in software engineering to seek out each other and start this important dialogue. Many of us have been challenged by the fact that educational research is not recognized as ‘‘legitimate” technical research. It somehow does not ‘‘count” when it comes to publications, conferences, tenure, and so on, except at those universities with a strong commitment to education as a legitimate area of research in its own right. We need to continue to push for recognition of educational research as a legitimate field of research. We need to further the profession. We need to recognize the need to be change agents and to motivate the next generation of software engineers and software engineering educators. Software engineering specializations within a broad variety of degree programs will help to reach this education goal. Professional certifications such as the IEEE Certified Software Development Professional (CSDP) as well as professional society memberships may help in this area. We are seeing specializations in related areas, such as information security, software security, and software assurance, as well as certifications in these areas by SANS and others. These are clearly aimed at security of our software and networks. We also need to embrace the IEEE Code of Ethics (IEEE, 2006) and use it to guide our work. These are big challenges, but I believe that if we work together as educators we can be successful in meeting the challenge. Our challenge is to exercise leadership by making things happen, not just by doing our job. It would not be easy, but it will be fun!
References Bagert, D.J., Mu, X., 2005. Current state of software engineering master’s degree programs in the United States. In: Proc. 35th Annu. Conf. Frontiers in Education Conference, IEEE Computer Society Press, New York, F1G-1–F1G-6.
575
Boykin, D., 2007. Is it time to license software engineers? PE Magazine.