Collaborative Model-Driven Software Engineering: a Classification Framework and a Research Map [ICSE 2018 - journal first]
2 likes854 views
The document presents a classification framework and research map for collaborative model-driven software engineering (CoMDE), emphasizing its growth due to factors like distributed agile and open-source development. It identifies challenges and publication trends in existing approaches, proposing a taxonomy to guide future research directions and highlight research gaps. Key future directions include better integration of collaborative dimensions, tools interoperability, and performance enhancements in CoMDE ecosystems.
![Ivano MalavoltaVRIJE
UNIVERSITEIT
AMSTERDAM
Collaborative Software Engineering - CoSE
”[...] any software project with more than one person is created through a process
of collaborative software engineering”*
What makes CoSE growing so much:
● distributed agile
● open-source
● global software development
* I. Mistrık, J. Grundy, A. Hoek, and J. Whitehead, Eds., Collaborative Software Engineering. Springer Berlin Heidelberg, 2010.
2](https://image.slidesharecdn.com/journalfirsticse2018-180604081429/85/Collaborative-Model-Driven-Software-Engineering-a-Classification-Framework-and-a-Research-Map-ICSE-2018-journal-first-2-320.jpg)
![Ivano MalavoltaVRIJE
UNIVERSITEIT
AMSTERDAM
Collaborative Software Engineering - CoSE
* I. Mistrık, J. Grundy, A. Hoek, and J. Whitehead, Eds., Collaborative Software Engineering. Springer Berlin Heidelberg, 2010.
Communication Coordination
Cooperation
Awarenessdemands
arranges
tasks for
mediates/
fosters
mediates/
fosters
mediates/
fosters
generates
commitments
3
”[...] any software project with more than one person is created through a process
of collaborative software engineering”*
What makes CoSE growing so much:
● distributed agile
● open-source
● global software development](https://image.slidesharecdn.com/journalfirsticse2018-180604081429/85/Collaborative-Model-Driven-Software-Engineering-a-Classification-Framework-and-a-Research-Map-ICSE-2018-journal-first-3-320.jpg)
Collaborative Model-Driven Software Engineering: a Classification Framework and a Research Map [ICSE 2018 - journal first]
- 1. Collaborative Model-Driven Software Engineering: a Classification Framework and a Research Map ICSE journal first paper Transactions on Software Engineering (TSE) Mirco Franzago - University of L’Aquila (Italy) Davide Di Ruscio - University of L’Aquila (Italy) Ivano Malavolta - Vrije Universiteit of Amsterdam (The Netherlands) Henry Muccini - University of L’Aquila (Italy) VRIJE UNIVERSITEIT AMSTERDAM
- 2. Ivano MalavoltaVRIJE UNIVERSITEIT AMSTERDAM Collaborative Software Engineering - CoSE ”[...] any software project with more than one person is created through a process of collaborative software engineering”* What makes CoSE growing so much: ● distributed agile ● open-source ● global software development * I. Mistrık, J. Grundy, A. Hoek, and J. Whitehead, Eds., Collaborative Software Engineering. Springer Berlin Heidelberg, 2010. 2
- 3. Ivano MalavoltaVRIJE UNIVERSITEIT AMSTERDAM Collaborative Software Engineering - CoSE * I. Mistrık, J. Grundy, A. Hoek, and J. Whitehead, Eds., Collaborative Software Engineering. Springer Berlin Heidelberg, 2010. Communication Coordination Cooperation Awarenessdemands arranges tasks for mediates/ fosters mediates/ fosters mediates/ fosters generates commitments 3 ”[...] any software project with more than one person is created through a process of collaborative software engineering”* What makes CoSE growing so much: ● distributed agile ● open-source ● global software development
- 4. Ivano MalavoltaVRIJE UNIVERSITEIT AMSTERDAM Collaborating on what? ... on the code ... on documents ... on design artifacts ... on project management artifacts This work focuses on specializing CoSE to Model-Driven Software Engineering (MDSE) 4
- 6. Ivano MalavoltaVRIJE UNIVERSITEIT AMSTERDAM What is Collaborative Model-Driven Software Engineering? 6 Pilot studies about CoSE+MDE Secondary studies about CoSE Analysis and open coding Preliminary definition of Collaborative MDE Discussions with experts Definition of Collaborative MDE
- 7. Ivano MalavoltaVRIJE UNIVERSITEIT AMSTERDAM Modelmanagement Collaboration Communication Models Collaborative MDSE approach. A method or technique in which multiple stakeholders manage, collaborate, and are aware of each others’ work on a set of shared models. 7 Collaborative Model-Driven Software Engineering
- 8. Ivano MalavoltaVRIJE UNIVERSITEIT AMSTERDAM Problem The body of scientific literature on Collaborative MDSE is LARGE and SCATTERED ● Identify, classify, and understand ● the publication trends, characteristics, and challenges ● of existing collaborative MDSE approaches ● from a researcher’s viewpoint GOAL 8
- 9. Ivano MalavoltaVRIJE UNIVERSITEIT AMSTERDAM Systematic mapping study What are the characteristics of collaborative MDSE approaches? What are the challenges and shortcomings of existing collaborative MDSE approaches? What are the publication trends that can be deduced from the scientific publications about collaborative MDSE approaches over time? RQ1 RQ2 RQ3 ACM Digital Library IEEE Explore ISI Web of Science Science Direct SpringerLink Wiley InterScience Initial search False Positives removal Selection criteria Clustering Snowballing 3,047 78 20 1,462 900 175 175 1,047 1,047 62,238 317 59 59 48+ (collaborat∗ OR coordinat∗ OR cooperat∗ OR concur∗ OR global) AND (MDE OR MDD OR MDA OR MDS∗ OR EMF OR DSL OR DSML OR ”model driven” OR ”eclipse modeling framework” OR ”domain specific language” OR ”domain specific modeling language”) 1,066 1,008 9
- 11. Ivano MalavoltaVRIJE UNIVERSITEIT AMSTERDAM How to use the taxonomy? As a classification framework ● for building a map of the state of the art ● for putting YOUR approach into context Taxonomy for Collaborative MDSE 1 11
- 12. Ivano MalavoltaVRIJE UNIVERSITEIT AMSTERDAM How to use the taxonomy – classification framework Examples 1 12
- 13. Ivano MalavoltaVRIJE UNIVERSITEIT AMSTERDAM How to use the taxonomy? As a classification framework ● for building a map of the state of the art ● for putting YOUR approach into context As a RADAR for research gaps ● for extracting current limitations & challenges ● for identifying future research directions Taxonomy for Collaborative MDSE 1 2 13
- 14. Ivano MalavoltaVRIJE UNIVERSITEIT AMSTERDAM What’s next? Main future directions ● Better integrate the 3 dimensions of CoMDSE ● Support for custom collaboration workflows ● Tools interoperability ○ modeling, analysis, simulation tools, ... ● Model synchronization with change propagation ○ Conflicts management ● Performance and scalability 14 M. Franzago, D. Di Ruscio, I. Malavolta, H. Muccini. “Envisioning the Future of Collaborative Model- Driven Software Engineering”. In Proceedings of the IEEE/ACM 39th International Conference on Software Engineering Companion (ICSE-C), 2017.
- 15. Ivano MalavoltaVRIJE UNIVERSITEIT AMSTERDAM What’s next? A proposal A Microservice-based ecosystem for Collaborative MDSE Ivano Malavolta VRIJE UNIVERSITEIT AMSTERDAM What’s next? Main future directions ● Better integrate the 3 dimensions of CoMDSE ● Conflicts management ● Model synchronization with change propagation ● Support for custom team collaboration workflows ● Tools interoperability ○ modeling, analysis, simulation tools, ... ● Performance and scalability • Highly customizable • Ecosystem of modeling artifacts and supporting services o Loosely coupled and independently deployable microservices • Distributed and light-weight model editors • Multiple languages and multiple views • Support for diverse stakeholders Our taxonomy as domain model for guiding the definition of the architecture of the ecosystem 3 15
- 16. Ivano MalavoltaVRIJE UNIVERSITEIT AMSTERDAM The CoMDSE ecosystem 16
- 17. Ivano MalavoltaVRIJE UNIVERSITEIT AMSTERDAM Conclusions Ivano MalavoltaVRIJE UNIVERSITEIT AMSTERDAM Systematic mapping study What are the characteristics of collaborative MDSE approaches? What are the challenges and shortcomings of existing collaborative MDSE approaches? What are the publication trends that can be deduced from the scientific publications about collaborative MDSE approaches over time? RQ1 RQ2 RQ3 ACM Digital Library IEEE Explore ISI Web of Science Science Direct SpringerLink Wiley InterScience Initial search False Positives removal Selection criteria Clustering Snowballing 3,047 78 20 1,462 900 175 175 1,047 1,047 62,238 317 59 59 106 (48) + (collaborat∗ OR coordinat∗ OR cooperat∗ OR concur∗ OR global) AND (MDE OR MDD OR MDA OR MDS∗ OR EMF OR DSL OR DSML OR ”model driven” OR ”eclipse modeling framework” OR ”domain specific language” OR ”domain specific modeling language”) 1,066 1,008 7 Ivano MalavoltaVRIJE UNIVERSITEIT AMSTERDAM Elicited taxonomy 8 Ivano MalavoltaVRIJE UNIVERSITEIT AMSTERDAM The CoMDSE ecosystem M. Franzago, D. Di Ruscio, I. Malavolta, H. Muccini. “Envisioning the Future of Collaborative Model-Driven Software Engineering”. In Proceedings of the IEEE/ACM 39th International Conference on Software Engineering Companion (ICSE-C), 2017. 3 14 Ivano MalavoltaVRIJE UNIVERSITEIT AMSTERDAM Collaborative Model-Driven Software Engineering Modelmanagement Collaboration Communication Models Collaborative MDSE approach. A method or technique in which multiple stakeholders manage, collaborate, and are aware of each others’ work on a set of shared models. 6 Replication package: http://www.di.univaq.it/mirco.franzago/collaborativeMDSE/
- 19. Ivano MalavoltaVRIJE UNIVERSITEIT AMSTERDAM Community activities Workshop @MODELS since 2016 Workshop paper submission deadline: July 17, 2018 IEEE Software Theme Issue ○ 31 Submitted papers ○ to appear on Nov/Dec 2018 http://cs.gssi.it/commitmde2018 19
- 20. Ivano MalavoltaVRIJE UNIVERSITEIT AMSTERDAM How to use the taxonomy – radar for research gaps Example 15 9 5 4 4 3 2 2 2 2 2 0 2 4 6 8 10 12 14 16 Conflicts management Model synchronization and change propagation Constraints specification, enforcing, and conformance checks Support for collaboration workflows and integration with development… Tool improvement Support for versioning and models editing history Support for combined asynchronous and synchronous collaboration Independence from any specific modeling language Interoperability with external tools Better coverage of modeling languages concepts Support for audio communication Identified limitations and shortcomings (as emerged from primary studies) 2 20
- 21. Ivano MalavoltaVRIJE UNIVERSITEIT AMSTERDAM Taxonomy – model management 21
- 22. Ivano MalavoltaVRIJE UNIVERSITEIT AMSTERDAM Taxonomy – collaboration 22
- 23. Ivano MalavoltaVRIJE UNIVERSITEIT AMSTERDAM Taxonomy – communication 23
- 24. Ivano MalavoltaVRIJE UNIVERSITEIT AMSTERDAM Findings about Model Management Editors are mostly desktop-based Prevalence of collaborative approaches supporting UML models (for any application domain) Collaborative model validation is only partially supported Multi-view support is quite limited Limited support to model versioning 24
- 25. Ivano MalavoltaVRIJE UNIVERSITEIT AMSTERDAM Findings about Collaboration Collaborative infrastructure is mostly based on a star-like topology with data stored in a central server Collaboration is mostly based on a remote shared workspace Versioning is mostly supported through a centralized versioning system architecture Branches creation is only sporadically supported while conflict detection is employed by over half of the approaches 25
- 26. Ivano MalavoltaVRIJE UNIVERSITEIT AMSTERDAM Findings about Communication Chat and annotations are the most used communication means Communication among technical stakeholders, with a unique approach designed for non-technical ones Workspace awareness - real-time updates and notification systems 26
- 27. Vision Model Management •Distributed modeling environment •Technical and non-technical stakeholders •Light-weight and multi-device modeling (mobility) •Large scale modeling (scalability!) •Cloud-based services, on-demand modeling, modeling- as-a-service •Extensibility/customizability
- 28. Vision Collaboration •Seamless sync and async modeling •Advanced inconsistency management •Model repositories with high performance and capacity (efficient model indexing and querying) •Advanced engines for detecting and resolving inconsistencies •Automation for branching and merging capabilities •Multi-view and multi-DSL support 28
- 29. Vision Communication •Improved usability and awareness •Advanced roles management •Design decision management and traceability •Integration with external communication means/tools (social, wiki, etc…) 29
- 30. Flexibility aspects •Support for importing, linking, and editing models conforming to several and integrated domain-specific modeling languages •Pluggable external editors to edit modelling artifacts (e.g., models, metamodels) •Support for including an open set of plugins providing vertical functionalities (e.g., for performing a specific type of analysis) •Traceability engines for external artifacts, possibly with links at the model element level •Concrete syntax redefinition or customization •Pluggable domain-specific conflict detection/resolution engines •Pluggable domain-specific consistency checkers, both within and across models 30