[Paper Review] Exploring software developers work practices

Sherlock A. Licorish, Stephen G. MacDonell|arXiv (Cornell University)|Apr 1, 2017

Software Engineering Techniques and Practices137 references18 citations

TL;DR

This study analyzes 30,646 work items from 474 IBM Jazz practitioners over three years to explore how diverse software development tasks—especially defect fixing—affect developer engagement and task resolution speed. It reveals that behavioral, intrinsic, and extrinsic factors collectively predict task resolution speed, with work nature significantly influencing engagement and performance.

ABSTRACT

We studied IBM Jazz practitioners interactions and task performance.This work mined logs from 30,646 work items (WIs) created since 3 years by 474 practitioners.Software practitioners studied were mostly involved in fixing defects.The nature of the work being performed influenced practitioners engagement.Combination of behavioral, intrinsic, and extrinsic factors affect task performance. In seeking to understand the processes enacted during software development, an increasing number of studies have mined software repositories. In particular, studies have endeavored to show how teams resolve software defects. Although much of this work has been useful, we contend that large-scale examinations across the range of activities that are commonly performed, beyond defect-related issues alone, would help us to more fully understand the reasons why defects occur as well as their consequences. More generally, these explorations would reveal how team processes occur during all software development efforts. We thus extend such studies by investigating how software practitioners work while undertaking the range of software tasks that are typically performed. Multiple forms of analyses of a longitudinal case study reveal that software practitioners were mostly involved in fixing defects, and that their engagement covaried depending on the nature of the work they were performing. Furthermore, multiple external factors affected speed of task resolution. Our outcomes suggest that behavioral and intrinsic issues may interact with extrinsic factors becoming significant predictors of the speed of software task resolution.

Motivation & Objective

• To understand the full spectrum of software development activities beyond defect resolution, including their impact on developer engagement and performance.
• To investigate how behavioral, intrinsic, and extrinsic factors interact to influence task resolution speed.
• To extend existing repository mining studies by examining longitudinal, multi-faceted development processes across a diverse set of tasks.
• To uncover patterns in practitioner behavior that correlate with task resolution efficiency and engagement levels.

Proposed method

• Longitudinal analysis of 30,646 work items (WIs) logged over three years from IBM Jazz practitioners.
• Mining of software repository logs to extract task types, resolution times, and interaction patterns.
• Application of multiple analytical forms—behavioral, intrinsic, and extrinsic factor modeling—to assess their influence on task performance.
• Covariation analysis of engagement levels with task nature to identify behavioral and contextual patterns.
• Use of statistical modeling to identify significant predictors of task resolution speed.
• Integration of qualitative insights from task categorization (e.g., defect fixing vs. feature development) into performance analysis.

Experimental results

Research questions

• RQ1 How does the nature of software development tasks influence practitioner engagement and task resolution speed?
• RQ2 To what extent do behavioral, intrinsic, and extrinsic factors interact to predict task resolution speed?
• RQ3 How do defect-related tasks compare to other development tasks in terms of engagement and performance?
• RQ4 What patterns in work item resolution emerge from longitudinal analysis of practitioner interactions?

Key findings

• Practitioners were predominantly engaged in defect fixing, which dominated the task workload across the studied period.
• Engagement levels varied significantly based on the nature of the task, with higher engagement observed in complex or high-impact activities.
• The combination of behavioral, intrinsic, and extrinsic factors emerged as significant predictors of task resolution speed.
• External factors such as team context and tooling environment influenced resolution speed, particularly when combined with behavioral traits.
• Defect resolution was not uniformly fast; resolution speed varied widely, suggesting unmodeled influences beyond task type.
• Intrinsic motivation and behavioral patterns interacted with external constraints to shape performance outcomes.