Proactive Quality Guidance for Model Evolution in Model Libraries - PowerPoint PPT Presentation

Proactive Quality Guidance for Model Evolution in Model Libraries Andreas Ganser, Horst Lichter, Alexander Roth, and Bernhard Rumpe Setting the Scene If You Take One Thing The Details Some References

Setting the Scene ... Model Recommenders and Model Libraries Do models change here? 2

If You Take One Thing ... Models Evolve in Model Libraries and Need Guidance How do they evolve? 3

Foundations: Evolving Models in Model Libraries Evolving Models Evolution Stages • Put model under monitoring • Vague • Review model and set quality gates • Decent • Resolve model issues and enhance it • Fine • Focus: evolution workflow support Goal: reusable, recommendable models 4

Foundations: Quality Stages, Gates, and Model 5

Foundations: Proactiveness and Guidance Foundation for Proactiveness Foundations for Guidance • Strong Attributes • Defect Defects Dangling references • • Syntax checker & metrics Missing names • • Checker • Not well formed • Smell • Medium Attributes Smells Too many classes • • Metrics & reviews Good class • • Thresholds Not well extracted • • Hunch • Weak Attributes Design contradicts content Hunches • • Design is awkward Reviews & judgement • • Thinking hats Not well designed • How to enable this? 6

Metrics and Simple Reviews Existing Metric Suites Simple Reviews • Use what‘s there ... • “Real” reviews too complex • Simplified reviews (streamlining) C&K Suite, Frankel, Genero, Martin, Ramirez, … • • Link to quality model • Idea: Six Thinking Hats become Five Review Hats • Yellow Hat (Good Points Judgment) • Black Hat (Bad Points Judgment) • White Hat (Information) • Green Hat (Creativity) • Red Hat (Emotions) • 7

Proactive Quality Guidance: A Software Prototype 8

Some References What else is going on ...? The HERMES Project References A. Ganser, H. Lichter, Engineering Model Recommender Foundations – From Class Completion to Model Recommendations, (Modelsward 2013, Spain) A. Ganser, T. N. Viet, H. Lichter, Multi Back-Ends for a Model Library Abstraction Layer, (ICCSA 2013, Vietnam) A. Dyck, A. Ganser, H. Lichter, Enabling Model Recommenders for Command- Enabled Editors, (MoDELS MDEBE 2013, US) and more to come on Model Recommender UI Survey, Framework Internals, Contexts / Scanners 9

Thanks for your attention … any questions? 10