Do static type systems improve the maintainability of software systems? An empirical study

Abstract

Abstract-Static type systems play an essential role in contemporary programming languages. Despite their importance, whether static type systems influence human software development capabilities remains an open question. One frequently mentioned argument for static type systems is that they improve the maintainability of software systems-an often used claim for which there is no empirical evidence. This paper describes an experiment which tests whether static type systems improve the maintainability of software systems. The results show rigorous empirical evidence that static type are indeed beneficial to these activities, except for fixing semantic errors.

An empirical study of the influence of static type systems on the usability of undocumented software

“…Finally, Kleinschmager et al performed an experiment on 33 subjects which combined repetitions of previous experiments (with Java and Groovy as languages) [15]. Among other tested hypotheses, it was confirmed that fixing type errors in a statically typed language is faster than corresponding fixes in a dynamically typed language, while no differences for fixing semantic errors was measured.…”

“…Consequently, using Eclipse would probably have been an advantage for Java developers and would have confounded the measurements: the intention of the experiment was to measure the impact of the language feature type system, not the maturity level of tool support. Hence, we decided to use a plain text editor with syntax highlighting, as has been done in previous experiments [15,27]. This custom editor permitted the user to compile programs, run applications, and execute test cases.…”

“…The study presented here is part of a larger experiment series that analyzes the impact of static type systems on software development (see [12]). So far, we performed four other experiments ( [10,15,27,28]). 3 The study by Hanenberg studied the impact of statically and dynamically typed programming languages to implement two programming tasks within approximately 27 hours [10].…”

“…In order to address the problem that different languages have different language features and that consequently differences in an experiment cannot be reduced to the factor type system, it is necessary to find two very similar languages. According to previous experiments (see [15,27,28]) we decided to use Java and Groovy [16]. While Groovy has a number of language features in addition to Java, it can also be used as a "dynamically typed Java": all type declarations of variables, fields and parameters can be annotated with the keyword def (without referring to the corresponding nominal type).…”

“…The experiment by Kleinschmager et al [15] should also be taken into account; there, for no single task a positive impact of the dynamic type system could be shown. Comparing the tasks of the experiment presented here and the one in [15], we think that (again) the difference lies in the complexity of the programming tasks.…”