Requirements for Student Programs in the Undergraduate Curriculum: How Much Is Enough?
by Lionel E. Deimel and Mark Pozefsky

When I began teaching in the Computer Science Department at North Carolina State University, I taught the second programming course a number of times before I taught the first one. Intrinsically, that course emphasized some of the finer points of programming that had not been of much concern in the first course. The experience intensified my concern for aspects of programs that went beyond whether the code worked (or seemed to). I expected programs that displayed good programming technique, that were adequately documented, and that behaved acceptably when I (or a student assistant) executed them with data—sometimes bad data—the students had not seen. Eventually, I brought these concerns to the introductory course.

My requirements were not always popular with students, and I think even my colleagues were skeptical, if not of my aspirations, then at least of the practicality of implementation. Enforcing standards and providing personalized feedback on programs were difficult in a 30-student class. Doing so was considerably more of a problem in the introductory course at NCSU, which, at the time, had well over a hundred students in each lecture section.

This 1979 SIGCSE Bulletin paper (citation), written with colleague Mark Pozefsky, makes the case for imposing high standards on student programs across the computer science curriculum, beginning with the first course. Although it discusses some implementation issues, it does not address the difficult problem of managing the enforcement of those standards through a staff of student assistants.

— LED, 5/31/2007