Lightweight Model-Driven Engineering
Download as PPTX, PDF11 likes3,441 views
The document discusses lightweight model-driven engineering (MDE), emphasizing its potential to enhance productivity, quality, and maintenance in software development. It critiques the adoption of MDE, noting its challenges and the need for user-centered approaches, and presents educational attempts to teach MDE effectively through practical applications. Furthermore, it highlights the importance of verification, tooling advancements, and the need for MDE to adapt and serve various domains to remain relevant and impactful in the software engineering field.
![Model
Rapid prototyping
Static analysis
Code generation
Automated testing
Refactoring/
Transformation
Documentation
[Illustration by Bernhard Rumpe]
MDE = Models as key elements in all aspects of
Software Engineering](https://image.slidesharecdn.com/lightweightmdeclean-170615022511/85/Lightweight-Model-Driven-Engineering-15-320.jpg)
Lightweight Model-Driven Engineering
- 1. Lightweight Model-Driven Engineering Jordi Cabot @softmodeling – jordicabot.com June 2017
- 2. About
- 3. SOM Research Lab Software runs the world. Models run the software
- 4. Nuestra misión Interested in the broad area of systems and software engineering, especially promoting the rigorous use of software models and engineering principles in all software engineering tasks. Flickr/clement127
- 6. Disclaimer
- 7. “As a writer, my job is to change your perspective and make you think outside the box.” - Mary Sage Nguyen
- 8. Based on my background But keynote used as an excuse to reflect about Algú tirant-se a la piscina
- 11. • Grady Booch – history of softwre engineering The entire history of software engineering is that of the rise in levels of abstraction - Grady Booch Goal: Write less and less code
- 12. Everybody is doing it
- 13. MDE is just the logical next step
- 15. Model Rapid prototyping Static analysis Code generation Automated testing Refactoring/ Transformation Documentation [Illustration by Bernhard Rumpe] MDE = Models as key elements in all aspects of Software Engineering
- 16. The MDE equation Models + Transformations = Software
- 18. What is a model 18 City of Nantes = “system” to be modeled A map is a model of this system Its legend is the grammar/ metamodel Different maps can focus on different “views” of the city
- 19. Not so different from languages MDE Grammarware MOF (metametamodel) UML (metamodel) ABank.uml EBNF.g Java.g MyProgram.java
- 21. MDE benefits +productivity +quality +Maintenance …. 2X – 8X 1.2X – 4X 80%
- 22. http://matt.might.net/articles/phd-school-in-pictures I’ll skip my over 500 “introductory” slides to MDE (but they are onine)
- 24. BUT he is saying this since 1985
- 25. Is MDE dead?
- 26. There is still money to be made on this market
- 27. Do you really think Agile has more market penetration?
- 28. MDE Adoption
- 29. • Asterix view : resistentes al modelado
- 31. To model, or not to model, this is the WRONG question - Shakespeare
- 32. What/when/how many models ? - Depends on: Size, Team, Domain, Complexity Real question
- 34. Even if MDE is not dead… We can do better Increase adoption Focus on what matters to users Not losing the perspective
- 35. MDE Research vs User needs & interests
- 36. Nés papites que el papa Oh great UML Spirit, please grant me the insight to valuably use the UML, the strength to challenge its misuse, and the wisdom to
- 38. UML Fever + UML prayer
- 39. Operación bikini
- 40. Revisiting MDE under a ligthweight perspective
- 42. Everything is a model
- 43. Models in real development processes
- 44. Agile Modeling
- 45. Collaboro: end-user driven DSL Development • Providing means to discuss about language elements • Overcoming technical barriers Participation • Suitable environment fostering end-user discussion • Facilitating voting and decision processes • Keeping traceability Collaboration
- 47. Discussing the abstract syntax
- 48. Discussing the concrete syntax
- 50. Lightweight learning of MDE
- 51. MDE4EDU
- 55. What can we learn from this? How to effectively teach MDE?
- 57. 1st Attempt: Students as MDE devs 45h MDE Foundations MDE core technologies DSLs M2M M2T Methodology and Infrastructure Exercises on building new MDE artefacts (e.g. a new modeling language with code- generation) Too complicated – surviving mode MDE as devs – unlikely role in practice
- 58. 2nd Attempt: Students as MDE users 45h MDE Foundations MDE core technologies Methodology and Infrastructure Case Study 1: Java – to – UML reverse engineering Caset Study 2: Code-generation of a CRUD web- based app CSs chosen to highlight the benefits of MDE
- 59. Better in theory – Blame the tools for the results Lots of installation problems Lack of optimization for the generated app Tool crashes and corrupted projects Lack of documentation Difficult to customize the code
- 60. Blame the chosen examples Maybe good for the database and back-end, not the front-end This can already be done with current programming frameworks Missing common pre-implemented patterns (login, social networks) Difficult trade-off. • Too expressive and people feel it’s a huge investment to learn • Too simple and people don’t perceive the benefits
- 62. • Start with a very compelling development scenario • Change the requirements late in the development • Use a repetitive scenario • Compile a set of reference examples to give a kickstart • Keep you target audience in mind For instructors • Document, document and document • Hide all underlying technical details (also for errors) • Offer a well-packaged and standalone installation • Keep up with the trends in the software industry • Favour trust over everything else For tool vendors
- 65. GAMIFICATION FRAMEWORK FOR END USERS - THE DSL
- 66. GAMIFICATION FOR END USERS - THE TOOL NOTIFICATION S SUBMIT FORMQUERY DEFINITIONS GAME STATISTICS https://github.com/SOM-Research/gamification-modeling-learning
- 68. 68 Original model 1st refinement nth refinement Model Transformation Model Transformation Source Code ... MDE-based software development process Errors in models breed and generate more errors in the code Why model verification
- 69. Even “trivial” models can be wrong Student EnrolsIn20..* Course Likes 1 5 1
- 70. Person name: string children 2 * + constraint : Nobody can be his own ancestor parent
- 71. 71 © AtlanMod - atlanmod-contact@mines-nantes.fr …. .No finite (and non-empty) solution
- 72. Consistent? Proven? Formalism / Logics • Dynamic Logic (KEY) • High-Order Logics (HOL-OCL) • Deductive database queries (CQC) • SAT (Alloy + UML2Alloy), … Translate Prove Deduce UML model Class diagram + OCL constraints Typical solutions
- 75. Our pragmatic approach: Bounded verification 75 Solution? Constraint Satisfaction Problem 1. Variables – basic types + struct/list 2. Domains – finite 3. Constraints – Prolog 4. Property Additional Constraint Translate Solve Deduce Property? + Model Constraints
- 76. Resolution of the CSP Define cardinality variables Constraints on cardinalities Assign cardinalities Define attribute variables Constraints on attributes Assign attributes Proof
- 77. ATLTest (testing of M2M) 7 7 White-box approach: Analyzing the ATL code to derive test models that “exercise” it
- 79. Eclipse, Papyrus… are great for our community OPEN SOURCE Part of eclipse POWERFUL STABLE AND RELIABLE WORLDWIDE USED INDUSTRIAL-GRADE
- 80. Moving to the web: Back-end
- 83. • Addressable resources – Eclipse: platform:/resource/project/Simpsons.xmi – Web API: https://example.com/rest/Family/Simpsons Uniform and Constrained Interface & Statelessness ‒ CRUD operations using HTTP verbs : POST, GET, PUT, DELETE PUT https://example.com/rest/Family/Simpsons/parents/Homer • Representation-oriented – JSON: {"family":{"address":"742 Evergreen Terrace",…}} – XML: <family><address>742 Evergreen Terrace</address></family> 83
- 84. NeoEMF
- 85. Moving to the web: Client-side
- 86. Towards Automatic Generation of Web-based Modeling Editors Manuel Wimmer, Irene Garrigós, and Sergio Firmenich ICWE 2017
- 87. JavaScript client 87 Existing solution: IDE-based Modeling Editor M2T Java code EuGENia GMFM2MM2T Stencil code JointJs Runtime GMF Runtime DSML Definition Our contribution: Web-based Modeling Editor P l a t f o r m P l a t f o r m Model-to-Text (M2T) transformation to generate Stencils for the JointJS platform from EuGENia annotated metamodels
- 89. Show me only what I need to see
- 90. Lightweight MDE in other fields
- 91. Governance of open source projects
- 92. Code Community
- 94. Governance of open source projects is not explicit
- 95. ******
- 96. Project myProject { Roles: Committers Deadlines: myDeadline : 7 days Rules: myMajorityRule : Majority { applied to Task when TaskReview people Committers range Present minVotes 3 deadline myDeadline } } All the proposals for new development tasks will be accepted or rejected in 7 days by the committers of the project. Verbalization
- 98. Lightweight MdE for open data
- 99. “the goal of the project is to make the promise of open data a reality by giving non-technical users tools they can use to find and compose the information they need”
- 100. ? API Discovery: What APIs / open data sources?
- 101. ? S1 S2 S1 S2 ¿Y cómo los puedo combinar?
- 102. JSON Discoverer – Extracting implicit models in JSON documents
- 103. JSON Discoverer
- 104. Open Data for All
- 107. Lightweight Knowledge to use MDE
- 108. Cognification: The application of knowledge to boost the performance and impact of a process
- 109. During the industrial revolution, every machine got an electrified version The next revolution is the cognificatinof everything via cheap access to specialized AIs
- 111. Cognifying MDE: 5 examples Modeling bot as virtual assistant Model inferences to discover schema of unstructured data A code generator that mimicks a company programming style A real-time model reviewer A morphing modeling tool that adapts to the user profile
- 113. PhD thesis – CEA + SOM PhD Thesis: CEA + SOM
- 114. Final Thoughts: Research on Lightweight MDE
- 115. Most research papers on MDE are completely irrelevant for (most) end-users Category Theory for MT ATL MapReduce Temporal UML
- 117. We live in a MDE multi-verse We need to develop domain- specific MDEs to better serve the needs of users in different domains
- 118. "Publishing something in the most advanced journal doesn't change the world … What changes the world is converting research to something that is available in the market, pleases customers and improves the quality of life“ – Tharman. Deputy PM. Singapore
- 120. GrandMDE at STAF’17 • Screenshot of the program