I have just completed a HEFCE survey about accreditation of degree courses by the British Computer Society (BCS) that has the aim of soliciting opinion about whether accreditation of courses can be used to enhance graduate employability. The details of the survey are unimportant but what is certainly not addressed in the survey is whether or not accreditation is worthwhile. It is assumed that accreditation is a good thing and that tweaking the process will improve employability. Unfortunately, I believe that the current accreditation process as adopted by the BCS stifles innovation and provides spurious respectability to degree courses that produce graduates that are simply not very good.
If we look at employability statistics, there is a close correlation between the input qualifications and graduate employment after 6 months. This is not even slightly surprising – employers want to hire the most able students and hence these graduates are the ones who have no problem in getting jobs. What is covered in their degree course is largely irrelevant so long as there is a sound fundamental understanding of programming, abstraction and algorithms. With this as a basis, these graduates can easily learn new technologies and will do so several times in their working life. Accreditation of the courses taken by these students is unimportant – I don’t like it because it tends to stifle innovation but, apart from the waste of academic time in the involvement in the accreditation process, it is not particularly damaging.
Not all students entering computer science and related degree courses come with high input qualifications. In fact, probably the majority of CS students are accepted into degree schemes with relatively low entry requirements. It would be invidious to single out particular institutions but the reality is that many universities are financially reliant on maintaining a large intake to their computer science courses so, to make up numbers, accept students without paying too much attention to whether or not their background and qualifications level is appropriate.
Now, I do not believe that computer science should be an elitist profession and that these students should be denied opportunities to study the subject. What these students need is a course that is tailored to their background where much more time is spent on fundamentals simply because it takes their students longer to master these basics of the discipline. My experience is that in some universities it is possible to graduate in computer science with a good honours degree but be unable to construct a non-trivial computer program. These students may well know about business skills or whatever the pet interests of course lecturers are but without the mastery of fundamentals, they simply cannot adapt to modern software engineering.
Unfortunately, a course that simply focused on fundamentals would find it it difficult to be accredited under the current process. There is a set of requirements (hoops) that degree schemes must meet, topics that are expected to be covered and universities have to jump through these hoops in order to be accredited. While the accreditation process is not quite ‘one size fits all’ , my experience is that there is an expectation from people on accreditation panels that certain topics will be covered and so universities simply don’t have the option of delivering the fundamentals of the discipline.
To be fair, the roots of this problem do not lie in the accreditation process but in the ‘marketing’ of courses by some institutions that believe that students are attracted to buzzwords rather than fundamentals. But accreditation exacerbates the problem because it only looks at courses, persists in the delusion that all universities are of a comparable standard and fails to look at the quality and skills of the graduating students. It focuses on process and the material covered rather than whether or not the graduates are actually any good.
The fundamental problem with BCS accreditation is that it has bought into the quality delusion that process is all. This stems from Deming’s process improvement approach in manufacturing but we don’t manufacture graduates and it isn’t simply a problem of getting the process of setting up machines correct. If accreditation is to have any real value, it has to look at the quality of the output from a course and employability is one metric that can be used to assess that quality. Some, perhaps most, courses will fail. But accreditation will then be meaningful and have some relevance to employers.
There is a further problems with the current approach to accreditation which I believe is harmful. The most active and enthusiastic academics and practitioners are repelled by the bureaucracy of the process and refuse to become involved. Hence, those involved tend to be rather conservative, their knowledge is often out of date and they find it difficult to understand some of the material in courses. They tend to favour what they know and universities that try to innovate and discard some less relevant material are sometimes criticised (I speak from personal experience here). Like all process-focused activities, those who don’t step out of line are favoured rather than innovators.
As it stands, I see no value in the current BCS accreditation process. Tweaking the process will not help – it needs root and branch redesign.
Of course, this won’t ever happen. The BCS is run by volunteers who have a vested interest in maintaining the existing system and who will be reluctant to antagonise the universities whose courses should not be accredited. Universities, especially those that offer sub-standard courses, will actively oppose change as it would reveal their own inadequacies. Most potential and graduating students see no value in BCS accreditation and couldn’t care less about whether or not a degree scheme is accredited. Employers are not fooled by accreditation and focus on making their own assessment of skills. They consider the BCS to be largely irrelevant (which is a pity) and they certainly won’t pressurise the BCS to change.
I have been a Fellow of the BCS for 25 years and a member for rather longer than that. I have worked in universities where the courses have been BCS accredited and in St Andrews, one of the few universities which has actively opted out of the accreditation process (a decision made before I joined the university). The computer science course at St Andrews comes top of the Guardian league tables and consistently appears in the top half dozen or so courses in other league tables.