What should the core achieve?

1 Aug 2010

The core courses of a computer science curriculum should equip an undergraduate with the mindset and a tool set to tackle software design problems. As for the former, interviewers, managers, and professors should expect three different attributes. First, a successful “graduate” of the coreIn a typical curriculum, the core includes two course on programming, two courses on discrete data structures and algorithms, and a course on hardware platforms. At Northeastern, it also includes a course on logic in programming. should have internalized the basic elements of the software design process and should use them routinely. That is, when confronted with a problem, the student should attempt to solve it in a systematic manner: gathering and organizing data; working through examples to clarify ambiguities and to get an idea about the solution; translating the organization of the problem data into an organization of the solution; tackling the solution proper; and testing the solution for correctness and other constraints.

Second, the student should always be aware of non-functional constraints on projects. The engineering of software systems does not happen in a vacuum. Constraints may also include economic constraints, such as the cost of a project. Our core courses cannot address the solution space for these non-technical constraints, but they should remind students of their existence. In addition to working functionally, software systems must satisfy a range of constraints, e.g., end-to-end performance, memory use, or energy consumption. When systems fail to satisfy constraints, the students must know that the first step is to conduct measurements to find “hot spots,” followed by a diagnosis step, and ideally resolved with the use of standard solutions (which are less expensive and easier to maintain than novel solutions).

Third, the student should have no fear of new programming languages and/or new programming environments. In the modern world of computing, any large project—or any new job—involves several programming languages, each of which comes with its own set of tools and its own environment. A graduate of the core courses must know that a solid familiarity with these tools is necessary to become a productive programmer and what it takes to acquire this level of familiarity.

As for the tool set, the undergraduate should know several programming languages—including a declarative database language—so that acquiring a new one doesn’t seem to be a large obstacle; the typical range of important data structures and algorithms; and the practical experience with some tools to find “hot spots” and the theoretical knowledge to analyze and resolve these problems.