Mining Treatment-Outcome Constructs from Sequential Software Engineering Data

Abstract

Many investigations in empirical software engineering look at sequences of data resulting from development or management processes. In this paper, we propose an analytical approach called the Gandhi-Washington Method (GWM) to investigate the impact of recurring events in software projects. GWM takes an encoding of events and activities provided by a software analyst as input. It uses regular expressions to automatically condense and summarize information and infer treatments. Relating the treatments to the outcome through statistical tests, treatment-outcome constructs are automatically mined from the data. The output of GWM is a set of treatment-outcome constructs. Each treatment in the set of mined constructs is significantly different from the other treatments considering the impact on the outcome and/or is structurally different from other treatments considering the sequence of events. We describe GWM and classes of problems to which GWM can be applied. We demonstrate the applicability of this method for empirical studies on sequences of file editing, code ownership, and release cycle time.Scenario #1: Project manager Alice is looking into ways to enhance the healthiness of the code developed by her team. The team doesn't have a unified way of committing, testing and reviewing the code. Some of the sub-project teams follow a commit-test-review process, while others apply review-test-commit. The rest does not follow either of these strategies to review their code. Alice wonders if the selection of one of these strategies affects the number of code bugs. To answer this question, she selects GWM and encodes each method for testing and reviewing a file by the sequence of testing (T), committing (C), and reviewing Experiment design

Release conventions of open‐source software: An exploratory study

“…Xia et al 27 proposed a model to predict the crashing releases, where they also showed that development metadata are effective in such predictions. Nayebi et al demonstrate that the events and activities of software analysts are applicable for empirical studies on release cycle time 28 . However, the above studies discuss development data to predict releases, but the management data are not utilized to explore releases.…”

“…Nayebi et al demonstrate that the events and activities of software analysts are applicable for empirical studies on release cycle time. 28 However, the above studies discuss development data to predict releases, but the management data are not utilized to explore releases. This study uses both management and development data to explore the concrete number of releases at a particular time.…”

Does the duration of rapid release cycles affect the bug handling activity?

“…They obtained feedback from five Eclipse Core maintainers who confirmed their findings. To determine whether the release cycle duration affects the app rating, Maleknaz et al [9] analyzed 6,003 mobile apps through the GWM method. They found seven unique release sequences that significantly affect the app rating.…”

“…We group releases in five categories of release duration, and treat the releases of each project as a sequence of release categories. This allows us to use the Gandhi-Washington Method (GWM) implemented by Nayebi et al [9]. The GWM tool allows us to test if there is a statistically significant difference in a given metric w.r.t different release cycle durations.…”