Enriching Tool Support for Model-Driven Software Development
- 1. Universidad de los Andes Departamento de Ingeniería de Sistemas y Computación Enriching tool support for model-driven software development David Fernando Méndez Acuña Iván Mauricio Sánchez Morcote Advisor: Dr. Rubby Casallas MASTER’S THESIS DEFENSE 1
- 2. AGENDA • Context • Challenges of the model-driven engineer • What do we need to overcome the challenges • Thesis objectives – Individual contributions • Solution strategy • Implementation • Validation • Conclusions 2
- 3. Context • Model-driven software development is an approach that proposes to use models as first-class entities for creating software products. 3
- 4. Model-driven engineer Context Modeler Model-driven solution T1 T2 T3 4 Problem model(Domain- specific) Application code (Platform-specific) solves
- 6. Willy Context Carolina Model Transformation Chain for UML 2 SQLCode UML RDB SQL Code SQL Tm2m Tm2m Tm2t It is not easy! There are a lot of challenges to overcome <<UML>> University <<SQL>> University DB Schema Example solves 6
- 7. Challenges of the model-driven engineer 7 Development artifacts manipulation Execution of repetitive tasks Construction of the IDE for the modeler
- 8. Challenges of the model-driven engineer Development artifacts manipulation 8 Example Willy Model Transformation Chain for UML 2 SQLCode University UML University RDB University SQL University SQL UML RDB SQL Metamodels Models Transformations
- 9. Challenges of the model-driven engineer Development artifacts manipulation Example University UML University RDB University SQL University SQL UML RDB SQL ? Models are manipulated as files There is not information about how models are related among them (relationship metadata)
- 10. Challenges of the model-driven engineer 10 Development artifacts manipulation Execution of repetitive tasks Construction of the IDE for the modeler
- 11. Challenges of the model-driven engineer Execution of repetitive tasks 11 Example University UML University RDB University SQL University SQL UML RDB SQL It is different in JAVA
- 12. Challenges of the model-driven engineer 12 Development artifacts manipulation Execution of repetitive tasks Construction of the IDE for the modeler
- 13. Challenges of the model-driven engineer Construction of the IDE for the modeler <<UML>> University <<SQL>> University DB Schema One single clic 13 Example Carolina
- 14. Challenges of the model-driven engineer Construction of the IDE for the modeler ECLIPSE IDEExtension 14 Ok! I must learn how to build eclipse plug-ins
- 15. 15
- 16. 16 How we can help Willy? David + Iván
- 17. State of the art 17 Development artifacts manipulation Execution of repetitive tasks Construction of the IDE for the modeler What has be done?
- 18. State of the art Atlas Model Management Architecture (AMMA) Megamodel as metadata repository Model-centric views 18
- 19. State of the art 19 Family of interoperable languages for model management (EPSILON) Languages for defining operations for model manipulation (e.g., merge, compare, match) EML Epsilon Merge Language
- 20. Model Management Framework Tool (MMFT) Graphical editor for defining interaction of models and operations State of the art 20 Merge Graphical Editor
- 21. State of the art Summary (challenge 1) Atlas Model Management Architecture (AMMA) Family of interoperable languages for model management (EPSILON) Model Management Framework Tool (MMFT) Megamodel as metadata repository Model-centric views Languages for defining operations for model manipulation (e.g., merge, compare, match) Graphical editor for defining interaction of models and operations 21 Predefined types of models. What if Willy does need new ones?
- 22. State of the art 22 Development artifacts manipulation Execution of repetitive tasks Construction of the IDE for the modeler What has be done?
- 23. State of the art ANT (XML-Based) Simple tasks Composite tasks 23 XML-Based o o
- 24. MWE2 (Proper syntax) Simple tasks Composite tasks State of the art 24 Proper syntax o o
- 25. ATLFlow (ATL-Based) Simple tasks Composite tasks State of the art 25 Focused on model transformation chains written in ATL t t
- 26. State of the art Summary (challenge 2) ANT (XML-Based) Simple tasks Composite tasks MWE2 (Proper syntax) Simple tasks Composite tasks ATLFlow (ATL-Based) Simple tasks Composite tasks 26 Predefined operations. What if Willy does need new types of tasks? He would have to implement the corresponding connectors
- 27. State of the art 27 Development artifacts manipulation Execution of repetitive tasks Construction of the IDE for the modeler What has be done?
- 28. State of the art Framework- specific modeling Framework-specific Eclipse-based 28
- 30. State of the art Summary (challenge 3) 30 Framework- specific modeling Framework-specific Eclipse-based Domain- specific tool generation Stakeholders-centric Eclipse-based Willy needs not only framework-specific but also domain-specific?
- 31. Solution strategy Overview 31 Development artifacts manipulation Megamodel as metadata repository Model-centric views DSL for definition of models interaction
- 32. Execution of repetitive tasks Solution strategy Overview 32 Event-action rules for tasks automation Support for composite tasks
- 33. Construction of the IDE for the modeler Solution strategy Overview 33 DSL for modeler’s IDE definition Automatic generation of modeler’s IDE
- 34. Solution strategy Overview 34 Development artifacts manipulation Megamodel as metadata repository Model-centric views DSL for definition of models interaction Execution of repetitive tasks Event-action rules for tasks automation Support for composite tasks Construction of the IDE for the modeler DSL for modeler’s IDE definition Automatic generation of modeler’s IDE
- 35. Thesis objectives Megamodel as metadata repository Model-centric views Models searching Simple tasks Composite tasks Framework-specific Eclipse-based Stakeholders-centric Challenge 1: Development artifacts manipulation Challenge 2: Execution of repetitive tasks Challenge 3: Construction of the IDE for the modeler 35
- 36. Individual contributions Megamodel as metadata repository Model-centric views Models searching Simple tasks Composite tasks Framework-specific Eclipse-based Stakeholders-centric Challenge 1: Development artifacts manipulation Challenge 2: Execution of repetitive tasks Challenge 3: Construction of the IDE for the modeler David Iván David + Iván 36
- 37. Solution strategy Model-driven engineer Modeler Model-driven solution Model management system Metadata Views Searching Execution rules MoDeTo MMS engineer Integrated Development Environment Problem model(Domain- specific) Application code (Platform-specific) One single clic MoMS (Domain-specific language) ? 37
- 39. Model Management System Implementation (Overview) MoMS Script M.A.M Generation process Modeling artifacts manager Operations launcher Modeler’s IDE generator Tools deposit MoMS Model O.L. Generation process M.T.G. Generation process 39
- 40. Implementation (Modeling Artifacts Manager) Model Management System Modeling artifacts manager Operations launcher Modeler tooling generator Metadata registry 40
- 41. Implementation (Modeling Artifacts Manager) Model Management System Modeling artifacts manager Operations launcher Modeler tooling generator Model Metamodel conformsTo Type Metamodel{ attributes{ attribute absolutePath : string } } Type Model{ attributes{ attribute absolutePath : string } references{ reference conformsTo : Metamodel } 41
- 42. Implementation (Operations Launcher) Model Management System Modeling artifacts manager Operations launcher Modeler tooling generator Model-driven engineer Operation Simple Composite 42
- 43. Implementation (Operations Launcher) Model Metamodel conformsTo validate OnUpdate OnUpdate operations{ simpleOperation validate( ){ Validator.validate(this.absolutePath, this.conformsTo.absolutePath); } } earules{ earule OnUpdate -> this -> this.validate(); earule OnUpdate -> this. conformsTo -> this.validate(); } } Model Management System Modeling artifacts manager Operations launcher Modeler tooling generator 43
- 44. Implementation (Modeler’s IDE generation) Modeler Input Model(Domain- specific) Application code (Platform-specific) One single clic 44
- 45. Implementation (Modeler’s IDE generation) Model Management System Modeling artifacts manager Operations launcher Modeler’s IDE generator Model-driven engineer MoDeTo 45
- 46. Validation MoMS script 1 MoMS script 2 Case study A: From UML 2 SQL Case study B: ViVaGUI Case study C: Maze Games Scenario 1: Simple MTCs Scenario 2: Weaving MTCs 46