International Conference on Software Engineering (ICSE’19) Socio-Technical Work-Rate Increase Associates With Changes in Work Patterns in Online Projects Farhana Sarker Bogdan Vasilescu Kelly Blincoe Vladimir Filkov Image source: h.p://clipart-library.com/cartoon-pulling-hair-out.html @b_vasilescu @KellyBlincoe @vlfilkov
High workload
High work-rate, high stress ❖ GitHub is like being onstage (Dabbish et al. 2012), (Marlow et al. 2013) ❖ Communication overload causes stress and reduces productivity (Reinke and Chamorro-Premuzic 2014), (Kalman and Rafaeli 2011), (Reinecke et al. 2017) ❖ Social pressure to respond quickly is associated with burnout and anxiety Image source: h.ps://octodex.github.com/ (Reinecke et al. 2017) ❖ Multi-tasking common, causes stress (Vasilescu et al., 2016), (Mark et al., 2008)
Social workload Prevalence of Perceived Causes and Effects of increased workload ? ? ? high work-rates Impacts of Stress on developers’ work patterns Sample: • 57K+ developers • 10+ years of activity • 150K+ repositories Image source: h.ps://octodex.github.com/ Data mining + Developer Survey
Developer survey Pilot Survey Survey 45 Responses (13%) 465 Responses (23%) Open-ended Questions Likert-scale questions My work-related stress is caused by What causes you work-related stress? deadlines. Strongly disagree Somewhat disagree deadlines Neither agree nor disagree Somewhat agree Strongly agree
Findings
High work-rates common Prevalence of ? high work-rates ● ● 40 1 Std. Dev. 1 Std. Dev. ● ● Mean Mean Number of Comments ● Number of Commits ● 20 30 ● ● ● ● 20 ● ● ● ● ● ● ● ● ● 10 ● ● ● ● ● ● ● ● ● ● ● 10 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ●● ● ● ● ● ● ● ● ● 0 0 2012 2014 2016 2012 2014 2016 Week Week
Causes of workplace stress Perceived Causes ? and Impacts of Stress Time pressure / too much work 9% 10% 81% Deadlines 10% 14% 76% Working on many things in parallel 15% 17% 68% Unclear or unrealistic requirements 18% 18% 65% Co-workers or manager 24% 22% 54% Performance / quality pressure 20% 28% 51% Communication issues 24% 31% 44% Unfulfilling work 35% 22% 44% ` Culture or language barriers 50% 33% 17% 100 50 0 50 100 Percentage Response Strongly disagree Somewhat disagree Neither agree nor disagree Somewhat agree Strongly agree
Impacts of stress Perceived Causes ? and Impacts of Stress I work outside of my normal business hours 11% 19% 70% the number of hours I work is 15% 17% 68% my communications are terse 10% 26% 64% my communications are negative 5% 39% 56% I stop working on the project 30% 38% 32% I am productive 52% 32% 16% I communicate 60% 25% 15% I produce high quality code 50% 45% 5% 100 50 0 50 100 Percentage Response Much less Somewhat less About the same Somewhat more Much more Senti4SD (Cafelato et al., 2018)
Empirical Study Effects of increased workload ? on developers’ work patterns Image source: h.ps://octodex.github.com/
Multitasking dimensions 1. Projects per day Working sequentially vs. Within-day multi-tasking A A B B Project Project C C D D 1 2 3 4 5 6 7 1 2 3 4 5 6 7 Day Day (Vasilescu et al., 2016)
Multitasking dimensions 2. Weekly focus Mostly on one project vs. Evenly on all projects 100% 100% High focus Low focus 80% 80% % Comments % Comments 60% 60% 40% 40% 20% 20% 0% 0% A B C D A B C D Project Project (Vasilescu et al., 2016)
Multitasking Dimensions 3. Day-to-day focus Repetitive day-to-day vs. Changing focus each day A A Project Project B B 1 2 3 4 5 6 7 1 2 3 4 5 6 7 Day Day (Vasilescu et al., 2016),(Xuan et al., 2014)
Statistical Analysis Responses Multi-tasking Predictors • Lines of code added A B • Comments C D • Sentiment A B C D 1 2 3 4 5 6 7 Controls A B • Total projects 1 2 3 4 5 6 7 • Time since first commit • Company affiliation • … Sample: • 57K+ developers • 10+ years of activity • 150K+ repositories
Effects: Weekly comments Day-to-day focus (less vs. predictable) 1 2 3 4 5 6 1 2 3 4 5 6 Company affiliation
Effects: Lines of Code added Higher LOC added High levels of comments vs. per day 1 2 3 4 5 6 7 1 2 3 4 5 6 7 Focused vs. commenting A B C D A B C D Day-to-day focus vs. (repeatability) 1 2 3 4 5 6 1 2 3 4 5 6
Effects: Negative sentiment ❖ Higher than normal levels of commenting Focused vs. commenting ❖ Commenting on many A B C D A B C D more projects than usual ❖ Committing on many Day-to-day more days per week focus vs. (repeatability) than usual 1 2 3 4 5 6 1 2 3 4 5 6
Thanks! Farhana Sarker Bogdan Vasilescu Kelly Blincoe Vladimir Filkov Image source: h.p://clipart-library.com/cartoon-pulling-hair-out.html
Recommend
More recommend