Model-Based Software Engineering: A Multiple-Case Study on Challenges and Development Efforts
- 1. Model-Based Software Engineering: A Multiple-Case Study on Challenges and Development Efforts Rodi Jolak Truong Ho-Quang Michel Chaudron Ramon Schiffelers MODELS, København- October 18, 2018
- 2. Motivation Empirical assessment of the use and process of MBE is scarce! The adoption of MBE is still debated in practice! 2 Productivity Quality Maintainability Time Consuming Unnecessary Effort Complicating Matters
- 3. Objective Address the discrepancy between belief and evidence in software engineering practices! 3
- 4. Research Questions • R.Q.1 How is the total effort spent on MBE distributed over different development activities? • R.Q.2 How large is the portion of collaborative work in MBE projects? • R.Q.3 What are the challenges that affect MBE in practice? 4
- 5. Case Study Design 5 Study Design Research Questions RQ1, RQ2, RQ3 Context Quantitative/Qualitative Data and Observations Case 1 MathWorks UoA 1 Effort Distribution UoA 2 MBE Challenges Context Case 2 PolarSys UoA 1 Effort Distribution UoA 2 MBE Challenges Quantitative/Qualitative Data and Observations Data Analysis and Interpretation of the Results
- 6. Context and Cases 6 Case 1 MathWorks Matlab & Simulink Case 2 PolarSys Papyrus & Capella 17
- 7. Organization of Development Teams 7 Team Leader Design Manager Test Manager Quality Manager Modelers/Developers Team Leader Design Manager Test Manager Quality Manager Modelers/Developers CASE 1: MathWorks CASE 2: PolarSys Project Manager Configuration & Integration
- 8. Data Collection 8 Online Questionnaire Procrasti Tracker Log Files
- 9. Data Analysis 9 Pattern Matching Cross-Case Synthesis
- 10. Results 10
- 11. 11 R.Q.1 How is the total effort spent on MBE distributed over different development activities? 0 2 4 6 8 10 12 14 16 Effort% MathWorks PolarSys
- 12. 12 R.Q.1 How is the total effort spent on MBE distributed over different development activities? Study Planning Requirements Specifications Design Coding Testing Integration Pressman Analysis & Design (40%) Coding (20%) Testing & Integration (40%) Sommerville Specs (15%) Design (25%) Development (20%) Testing (40%) Papatheocharous Plan (9.6%) Specs (9.3%) Design (14.0%) Build (42.3%) Testing (18.2%) Implement (6.6%) Heijstek Planning (13%) RE (8%) Analysis & Design (11%) Coding (38%) Testing (12%) Configuration (4%) Case Study 1 (MathWorks) Planning (15.0%) RE (10.0%) Analysis & Research (7.0%) Design & Modeling (17.0%) Code Generation (9.2%) Manual Coding (13.6%) Testing (15.4%) Integration & Configuration (12.7%) Case Study 2 (PolarSys) Planning (15.3%) RE (6.8%) Analysis & Research (8.8%) Design & Modeling (13.0%) Code Generation (12.0%) Manual Coding (11.0%) Testing (20.0%) Integration & Configuration (13.2%) 49% 23% 28% 44% 23% 33%
- 13. Implication 1: "MBE approaches do not require a lot of effort on design and modeling” “MBE approaches require little effort on manual coding, as most of the code is obtained from models via code-generation” 13
- 14. 14 How is the effort spent on different MBE development activities distributed over time? How is the effort spent on different MBE development activities distributed over time?
- 15. Implication 2: “Considering code-generation, we found that the tools offered by PolarSys open source technologies (i.e., Papyrus and Papyrus-RT) are more mature than the tools offered by MathWorks technologies (i.e., Matlab and Simulink).” 15
- 16. 16 R.Q.2 How large is the portion of collaborative work in MBE projects?
- 17. Implication 3: “Model-based software development is an endeavor that requires intensive communication and collaboration between developers.” 17
- 18. 18 R.Q.3 What are the challenges that affect MBE in practice? 0 5 10 15 20 25 Percentage% MathWorks PolarSys
- 19. Implication 4: “Our findings show that tool-related challenges are the most encountered. These tool-challenges are due to: tools usability, tool-chain learning, interoperability of tools, and tools installation and configuration.” 19
- 20. 20 How are the challenges that affect MBE distributed over project time?
- 21. Recap 21 Development Effort • MBE approaches do not require a lot of effort on design and modeling. • MBE is an endeavor that requires intensive communication and collaboration between developers. MBE Challenges • Tool-related challenges are the most encountered. These tool-challenges are due to: tools usability, tool- chain learning, interoperability of tools, and tools installation and configuration.
- 23. Future Work (2/2) 23 The Effect of Using Models on Software Design Communication How the Analyst described it How the Developer understood it What the Customer really needed Active Discussion Understand-ability Explaining Re-call
- 24. Thank You For Your Attention! Rodi Jolak* Truong Ho-Quang Michel Chaudron Ramon Schiffelers * www.rodijolak.com Development Effort • MBE approaches do not require a lot of effort on design and modeling. • MBE is an endeavor that requires intensive communication and collaboration between developers. MBE Challenges • Tool-related challenges are the most encountered. These tool-challenges are due to: tools usability, tool- chain learning, interoperability of tools, and tools installation and configuration.
- 25. Appendix (1/2) What were the discussions about? 25 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 1 2 3 4 5 6 7 8 Effort Week Discussion CASE 1: MathWorks Design/Development Testing Configuration and Integration Project Management 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Effort Discussion CASE 2: PolarSys Design/Development Testing Configuration and Integration Project Management