Is it easier to write matrix manipulation programs visually or textually? An empirical study

Empirical studies comparing the effectiveness of visual languages versus textual languages have rarely been attempted. We describe an experiment conducted on programmers solving vector and matrix manipulation tasks using the visual language Forms/3, the textual language Pascal, and a textual matrix manipulation language with the capabilities of APL. Our motivation, experimental approach, some of the difficulties experienced in attempting this type of empirical study, and a summary of the experimental results and insights gained are presented