Using Temporal and Semantic Developer-Level Information to Predict Maintenance Activity Profiles

Abstract

Abstract-Predictive models for software projects' characteristics have been traditionally based on project-level metrics, employing only little developer-level information, or none at all. In this work we suggest novel metrics that capture temporal and semantic developer-level information collected on a per developer basis. To address the scalability challenges involved in computing these metrics for each and every developer for a large number of source code repositories, we have built a designated repository mining platform. This platform was used to create a metrics dataset based on processing nearly 1000 highly popular open source GitHub repositories, consisting of 147 million LOC, and maintained by 30,000 developers. The computed metrics were then employed to predict the corrective, perfective, and adaptive maintenance activity profiles identified in previous works. Our results show both strong correlation and promising predictive power with R 2 values of 0.83, 0.64, and 0.75. We also show how these results may help project managers to detect anomalies in the development process and to build better development teams. In addition, the platform we built has the potential to yield further predictive models leveraging developer-level metrics at scale.

Boosting Automatic Commit Classification Into Maintenance Activities By Utilizing Source Code Changes

“…Its Kappa on the other hand, would be 0, making this model much less appealing. (3) Our previous work [5] shows that source code change types as defined by Fluri et al [11] are statistically significant in the context of maintenance activity categories defined by Mockus et al [1]. We believe that boosting (i.e.…”

“…First we classified the test dataset (the 15% of the entire labeled dataset) using a naive method to set an initial baseline. The naive method is based solely on searching for pre-defined words gathered from previous work [5], and returning the most frequent class (i.e., corrective) in case none of the keywords were present in a commit's message, see table 2 for more details. The results showed that 34.8% of the commits in the test dataset (60 commits) did not have any of the keywords present in their commit message, and were therefore automatically classified corrective.…”

“…Understanding these maintenance activities, performed in a source code repository, could help practitioners reduce uncertainty and improve cost-effectiveness [2] by planning ahead and preallocating resources towards source code maintenance. Maintenance activity profiles of software projects have therefore been a subject of research in numerous works [1][2][3][4][5][6]. To determine maintenance activity profiles, one must first classify the activities, which come in the form of developer commits to the version control system (VCS).…”

Processing Large Datasets of Fined Grained Source Code Changes

“…In the course of our studies [6][7][8], the data processing stage typically included the following aggregations: commit level; developer level; project level; global statistics. The analytical layer we present allows researchers to produce commit level aggregations (see Listing 5) and obtain statistics such as: change type frequencies, number of test case (test method) addition/removal/modification, number of test suite (test class) addition/removal/modification, associated ticket id, number of test files, and non test files in a given commit.…”

“…To effectively process large datasets, our analytical layer leverages Apache Spark [12] (henceforth Spark), a widely popular distributed computation engine. The analytical layer we suggest has been successfully used to conduct a number of studies in the field of software maintenance and evolution [6][7][8]. This leads us to believe it can be useful for researchers conducting studies that involve fine-grained source code changes.…”