2nd Workshop on Data and Models: Towards Intelligent Modelling Environments
The 2nd 1-Week Workshop on Data and Models at the Bellairs Research Institute of McGill University in Holetown, Barbados, is taking place from Friday, February 7th, 2020 to Friday, February 14th, 2020 with a special emphasis on Intelligent Modelling Environments. In essence, we want to gather ~15 researchers that are using (runtime) data and models in conjunction to investigate their interplay and complementarity for improving software development and maintenance tasks. To achieve such a purpose, the workshop will gather participants from various Software Engineering disciplines such as Model-Driven Engineering, Software Language Engineering, and Adaptive Systems.
Workshop Details
While the use of models has a longstanding history in software engineering practices, the use of (runtime) data becomes more and more prevalent in modern practices thanks to their availability, advances in data processing techniques, and the availability of resources on demand (e.g., cloud infrastructures). From existing search-based approaches to advanced predictive methods such as machine learning, the duality and complementarity of data and models appear more and more apparent in the various phases of the software development lifecycle. Such practices also open many opportunities in the development of future software systems for exploiting real data or engineering expertise as part of the whole software lifecycle.
We will focus on techniques that assist a modeller that interacts with a tool to build a model m by exploiting any kinds of data or knowledge:
- found in m (e.g., provide model completion to ensure coherence within the model),
- found in models related to m (e.g., provide modelling suggestions to ensure consistency with other views in a multi-view modelling approach),
- found in models of the same project (e.g., provide model completion to align m with the other models in the project),
- found in a library of reusable models, (e.g., suggest to apply reusable models found by mining for patterns),
- found in version control systems, (e.g., suggest to apply patterns found in previous projects),
- from runtime information (e.g., suggest updates to the model to evolve the model to correspond more closely to the real world, or give predictive information in the modelling tools about operational data such as resources consumption).
Assistance to the modeller will be investigated in a broad sense, from model auto-completion tools and modelling chatbots to literate and live modelling.
We expect the workshop to be highly interactive and open-minded, with the main objective to brainstorm on the opportunities and underlying challenges of the duality and complementarity of data and models in software engineering. As a possible outcome, we envision the collaborative definition of a roadmap on the challenges of intelligent modelling environments that would help the software engineering community to drive and structure future investigations.