Scrum2Kanban: Integrating Kanban and Scrum in a University Software Engineering Capstone Course SEEM’18 @ ICSE’18, Gothenburg, Sweden June 2018 Christoph Matthies [email protected] Enterprise Platform and Integration Concepts Hasso Plattner Institute, University of Potsdam 

Background Undergraduate Agile SE Capstone Course 2 

Background Undergraduate Agile SE Capstone Course w/ Kanban 3 

Research Questions 4 ■ What are students’ perceptions of Kanban practices? ■ Are those perceptions accurate? ■ How does using Kanban influence workflows? 

Method Two different approaches 5 Survey (N=18) Development Artifact Analysis (GitHub tickets & commits) 1 2 

Survey General Attitude 6 ■ Positive attitudes towards including Kanban ■ Recommended ■ Good understanding of agile ■ Preferred over last Scrum week ■ Neutral towards additional lectures, high variance Sanity Check 

Survey Extract of Attitudes towards Kanban 7 ■ Advantages of Kanban 

Survey Extract of Attitudes towards Kanban 8 ■ Advantages of Kanban ■ Drawbacks of Kanban 

Survey Extract of Attitudes towards Kanban 9 ■ Advantages of Kanban ■ Drawbacks of Kanban ■ Change in User Stories / Requirements 

Development Data Analysis Artifacts 10 ■ Length (title, body) ■ # Comments ■ # Interactions ■ Opened/Closed by ■ Assignee Commit History User Stories ■ Count ■ Files changed ■ Insertions ■ Deletions ■ Merge? 

Findings User Stories 11 User Stories were shorter when using Kanban ■ Mean body length was lower (513 vs. 367 chars), but not titles ■ Support for perception of US contents More dynamic interaction with US during Kanban ■ Only ~⅔ of user stories created by POs (vs 85%+ in Scrum) ■ Support for perception of autonomy Uneven task distribution ■ Not fixed by Kanban, # unique assignees did not significantly change ■ Support for perception ■ Identified need for improvement 

12 More commits ■ More non-merge commits (138 vs 289) ■ Support for hypothesis Smaller commits ■ Diff sizes similar ■ Hypothesis not validated Problem with Merges in Kanban ■ Mean amount of merge commits per week almost tripled (52 to 142) ■ Support for perceptions ■ Need for improvement Findings Commits 

Conclusions 13 ■ Students’ software development data ■ Another dimension of analysis ■ Addition to surveys ■ Artifacts in SE always produced, already there ■ Not everyone fills out voluntary survey 

14 ■ Students’ software development data ■ Another dimension of analysis ■ Addition to surveys ■ Artifacts in SE always produced, already there ■ Not everyone fills out voluntary survey ■ Contrasting perceptions and data can reveal areas of improvement / further research Conclusions 

Summary 

Image Credits 16 ■ HPI Campus by Stephan Schultz (CC BY 2.0) ■ Survey by Vectors Market from the Noun Project (CC BY 3.0) ■ analysis by Alvaro Cabrera from the Noun Project (CC By 3.0) ■ Service Report by Sophia Bai from the Noun Project (CC BY 3.0) ■ Merge by Danil Polshin from the Noun Project (CC BY 3.0) ■ GitHub Mark by GitHub Inc. (https://github.com/logos)