An approach to experimental evaluation of software understandability

A Model for Measuring Software Understandability

“…The experimental results of Figure 2 [3] show that the number of attempts to reconstruct and the number of components are correlated, so, the number of components can quantify understandability of software structure. Basing on literature [3], we can design membership function of understandability of software structure as follows: …”

“…The average number of attempts needed for correct reconstruction that one worker reconstructed many software systems means comprehension of the worker. The average number of attempts needed for correct reconstruction that many workers reconstructed one software system means understandability of the software system [3].…”

“…Basing on literature [3], we can design membership function of understandability of software structure as follows: …”

“…If it is difficult to understand, changes to it may cause serious faults and a chain reaction of changes. Such changes may cost more time than remaking the software system [3]. Figure 1 shows communications among humans, software, and hardware in software evolution.…”

“…Therefore, we usually consider that programmer 2 understands version 1 when he/she could correctly write version 2. In order to measure understandability, we need to observe an external process (like writing the version 2) which externalize understanding [3]. This paper proposes "integral of understandability" as a model for software understandability.…”