Human Error In Programming: The Need To Study The Individual Programmer
John E. Cooke, Richard B. Bunt · INFOR Information Systems and Operational Research · 1975
This paper discusses the problem of human error in programming, and suggests the importance of empirical validation of the efifectiveness of new styles or techniques, such as those generally considered as the components of “structured programming.” As a complement to studies of groups of programmers, it is suggested that studies are needed which concentrate on the information- processing limitations of human programmers.Following a review of relevant literature from other disciplines, eye-movements, perception and short-term memory are identified as important to the problem of reading and understanding a program. A possible experimental approach that might answer a number of current questions about program structure is suggested.