Key points are not available for this paper at this time.
An experiment was conducted to investigate how different types of modularization and comments are related to programmers' ability to understand programs. Forty-eight experienced programmers were given eight different versions of the same program and asked to answer a twenty question quiz used to measure comprehension. These eight different versions were the result of the program being constructed with four types of modularization (monolithic, functional, super, and abstract data type), each with and without comments. Those subjects whose programs contained comments were able to answer more questions than those without comments. Also, those subjects who were given the abstract data type version of the program were able to do significantly better than those with any other type of modularization.
Woodfield et al. (Mon,) studied this question.