Command-line usage in a programming exam

Introductory computer science education has a strong emphasis on the teaching and learning of programming skills. In establishing that a desired level of proficiency in the use of these skills has been attained, many courses implement a practical exam where students must complete a program and get it to run correctly under supervision and unaided. This exam may, as in our case, be presented as a "barrier" exam which must be passed in order to proceed to intermediate computer science enrolment. The importance of such an event is not always matched by our understanding (as educators) of student behaviour under such conditions. The binary (pass/fail) nature of the exam tends to be reflected in student perceptions of the exam, often polarised as being "quite easy" by those who pass or "too difficult" by those who fail. This paper describes an exploration into command-line behaviour during the exam, in an attempt to gain some insight into student behavioural reaction to it. In so doing, the issue is raised as to whether or not certain actions are more likely to serve as indicators of a successful candidate, and what meaning this has for teachers and students.