1st International Workshop on Debugging in Model-Driven Engineering

MDEbug 2017


Software Systems



===== Context =====
MDEbug is the first workshop focused on debugging in model driven engineering. Debugging can be defined as “locating the source of an observable defect”. A system contains a defect if it does not behave according to its requirements. A defect is typically observed by a V&V technique, such as testing, runtime verification, model checking, etc. When these techniques reveal that the system contains a defect, its source needs to be located. In software engineering, debugging is typically accomplished using techniques that inspect and manipulate an execution trace that led up to a defect. When models are used to develop systems, however, developers often have to resort to ad-hoc methods to debug the system. A common technique consists of inspecting/debugging the code generated from models. This is not ideal, since the developer has to switch contexts and is required to understand the semantics of the underlying implementation language. Moreover, the relation of the generated code to the higher-level modelling concepts is often not clear; this can be compared to debugging a program using an assembler debugger. With the growing importance of model-driven engineering techniques to develop complex systems, researchers are increasing the reliability of systems by, among others, integrating verification and validation techniques. This workshop aims to address a related issue, by investigating how program debugging techniques can be transposed onto the modelling realm and whether new, specific ones need to be developed. In particular, model debugging techniques need to take into account the wide variety of models (in a wide variety of languages) that are produced during system development.
Scope and Topics
In order to discuss these and further similar questions, we would like to invite submissions in the form of regular papers, position/experience papers, and demonstration papers (about novel tool features) related to the following topics:
- Debugging for languages with diverse semantics, including:
- non-determinism (e.g., Petri nets);
- concurrency (e.g., statecharts);
- continuous-time/discrete-time/discrete-event semantics;
- spatial distribution (e.g., cellular automata);
- dynamic structure (e.g., dynamic-structure DEVS);
- Debugging heterogeneous systems (modelled using multiple different modelling languages).
- Debugging for languages whose semantics are defined:
- operationally by building a simulator, executor, or interpreter;
- denotationally (or translationally), by mapping onto a domain with known semantics by building a model transformation, code generator, or compiler.
- Debugging for domain-specific languages.
- Tool support for model debugging.
- Techniques for “live modelling” (analogous to “live programming”).
- Debugging for model transformations.
- Techniques for omniscient/time-travel/reverse debugging.
- Debugging languages and their interpreters.
-Tracing support for modelling languages.
- Debugging of (instrumented) deployed systems, with feedback to the models that describe its design.
===== Workshop Format =====
MDEbug is a full-day workshop held as part of MODELS 2017. Since debugging in model-driven engineering is a relatively new research field with a growing community, this workshop’s main intention is to bring together interested researchers to focus the research effort. One objective will be to define the scope of debugging within modelling, and work out a (preliminary) roadmap for future research. The intent is to organize a highly interactive workshop, by stimulating discussion among researchers. We plan to have one keynote talk in the morning, followed by paper presentations. Afterwards, a number of discussion points will be decided upon and discussed in small, focused groups. We encourage submission of position papers that can stimulate the discussions, as well as more mature work on debugging.
===== Submission =====
Submitted papers should belong to one of these paper categories, depending on the nature of the contribution:
- Full research papers (up to 6 pages) which present novel innovative approaches.
- Position papers (up to 2 pages) which present new ideas or early-stage research.
- Experience reports (up to 6 pages) which present applications of an approach and extensively discusses the experiences of the researchers.
- Tool demo papers (up to 3 pages) which demonstrate a specific tool or extension to an existing tool that implement debugging techniques for models.
All submissions should follow the IEEE formatting guidelines: https://www.ieee.org/conferences_events/conferences/publishing/templates.html
Submit your papers electronically via Easychair: https://easychair.org/conferences/?conf=mdebug2017