Software Engineering at the Speed of Light: How Developers Stay - PowerPoint PPT Presentation

Software Engineering at the Speed of Light: How Developers Stay Current using Twitter Presented by: Brian Carignan

Introduction � Goal � Provide empirical evidence for benefits/drawbacks of Twitter � Reasons for use or non-use � Potentially improve tools � Qualitative study

Study Design � All participants recruited from GitHub � Grounded theory � Exploratory survey � 271/1160 (23%) responded � Open ended � Guide for interviews � Semi-structured interviews � Coding and memos � Validation survey � 1207/10000 (12%) responded � 677 said they used Twitter in their work

Research Questions 1. How does Twitter increase software developer awareness of people, trends, and practices? 2. How does Twitter help software developers extend their software knowledge ? 3. How does Twitter nurture relationships between software developers? 4. What are the challenges faced by software developers using Twitter, and how do they cope with them? 5. What are reasons for non-adoption of Twitter by software developers?

Findings

RQ1: Awareness � Activity � Following � Developers and projects � Technical news curators � Thought leaders � Activity � Promoting projects � Impact � Just in time awareness � Access to diverse opinions � Dissemination of knowledge � Increased adoption

RQ2: Knowledge � Activity � Asking and answering questions � Following experts � Participating in conversation � Impact � Learning as investment � Learning serendipitously � Undirected, not actively looking

RQ3: Relationships � Activity � Maintaining one’s image � Persona � Building communities � Hashtags � Impact � Discovering interesting developers � Achieving trust � Work validation � Feedback on projects � Formation of communities of practice � Job Opportunites

RO4: Challenges � Maintaining a relevant network � Find relevant developers � Unfollow developers � “Trial” � Consuming Content � Filtering � Skimming by profile pictures � Skimming often, read later � Reading routine

RO5: Reasons for Non-Adoption � Too much noise � 140-character constraint � Poor support for conversation � Unsure of benefits

Questions � Population representative of all developers? � GitHub, Twitter vs Private � Are the findings transferable to other platforms?