S. Winberg

University of Cape Town, Department of Electrical Engineering (SOUTH AFRICA)
In the area of computer engineering, the demand on graduates is becoming increasingly difficult to satisfy. It is not just too few graduates, but more their preparedness for industry and postgraduate work that is the difficulty. While some of the challenges identified in this paper are likely relevant to similar industry-university partnerships in other countries, this paper reports on an on-going partnership in South Africa. This partnership was a response to the Square Kilometer Array radio telescope that South Africa is bidding for together with eight other African nations. Through this bid, and South Africa’s broader investments into scarce skills, opportunities for funding have arisen, some of which involve industry-university partnerships. The partnership discussed in this paper is a three-way one involving two industry-based organizations and a university, respectively: the MeerKAT Project, the Centre for High Performance Computing (CHPC), and the University of Cape Town (UCT). This collaboration focuses on high-tech skills required by computer engineering graduates.

A computer engineer, as the name suggests, is an electrical engineer specialized in the design of computers. Traditionally, students in this discipline acquire a skill set that combines expertise from the fields of digital microcomputer design, electronics (e.g., building circuits) and computer science (e.g., algorithm design and programming). But, the landscape of computer systems is changing. This is partly due to programs and computer architectures having to be made more parallel to operate faster, since a ‘brick wall’ in terms of speeding-up individual CPUs has been reached.

In response to these changes in traditions, and the broadening variety of computer technologies, it is becoming difficult to deliver industry-ready graduates. In terms of the MeerKAT-CHPC-UCT partnership, this problem is fast becoming a critical one. While the accreditation process for a computer engineering degree imposes necessary constraints on the curriculum, the space in the curriculum for elective courses to prepare students for these local industries is barely sufficient – despite pressures to deliver well-prepared students according to the country’s skills development agenda. This problem is evident in delivering graduates to our MeerKAT and CHPC partners. Both partners need a similar set of common skills (mostly covered by the curriculum), but also additional skills. For example, CHPC needs graduates skilled in designing multi-threaded clustered computers; whereas MeetKAT needs experience in FPGA reconfigurable computers and software defined radio. These skills can take a year of courses to learn.

The academics in this partnership have reached a conundrum, which we colloquially call the ‘bludgeons or slingshots decision’: Do we want computer engineers that are traditionally skilled and can slowly ‘bludgeon’ their way in an inelegant Goliath manner to a result our industry partners could use? Or would we prefer more variety, as in graduates that each take aim with their individual ‘slingshots’ (i.e., special skills) and, like David, quickly and elegantly dispatch the problem concerned.

While a slingshot approach would be advantageous to our industry partners, we settled on a compromise. This paper proceeds by elaborating the specifics of the approaches we are using, which currently satisfies the accreditation requirements and makes provision for the industry partnership.