Summary of Workshop on Model-Driven Engineering Tools (MDETools’17) Mojtaba Bagherzadeh Francis Bordeleau Jean-Michel Bruel Juergen Dingel School of Computing Cmind Inc. IRIT School of Computing Queen’s University Ottawa, Canada University of Toulouse Queen’s University Kingston, Canada francis.bordeleau@cmind.io Toulouse, France Kingston, Canada mojtaba@cs.queensu.ca bruel@irit.fr dingel@cs.queensu.ca Sébastien Gérard Nicolas Hili Sebastian Voss CEA list School of Computing fortiss GmbH Saclay, France Queen’s University Munich, Germany sebastien.gerard@cea.fr Kingston, Canada voss@fortiss.org hili@cs.queensu.ca Abstract—The first workshop specifically devoted to tools III. P ROGRAM supporting Model Driven Engineering was held September 19, 2017 in Austin, Texas, USA. The motivation, scope, objectives, A total of eight submissions were received. Five of these and results of the workshop are summarized. were accepted. The program consisted of a keynote, paper Index Terms—Model-driven Engineering, tools, software and systems modeling presentations, a demo session, and a discussion session. The keynote ‘MDE Tools in Industry and Education: Under- I. M OTIVATION standing, Comparing and Improving the Tools’ was given by The easy availability of high-quality tools with effective Cortland Starrett, currently president of One Fact Inc, a com- supporting materials and documentation significantly increases pany developing open source modeling tools (BridgePoint) the chances of adoption for any new software development as well as modeling client applications. In his presentation, approach. Several research communities have recognized the Cortland drew on his experience in both industry and education importance of tools and, e.g., created workshops specifically and discussed some of the challenges of comparing tools, but designed to facilitate the evaluation and comparison of tools also showed several inspiring examples involving modeling (for, e.g., language workbenches [1], transformations [2], and design challenges such as ET-Robocon, a UML robot satisfiability solving [3], and verification [4]). contest that has been held annually since 2002 and whose In contrast, the modeling research community does not attendance has grown from 20 (2002) to 1800 (2013) with appear to be paying as much attention to effectively leveraging over 360 teams from industry, academia, and education [7]. tools for illustrating, evaluating, and disseminating research Two of the accepted papers used the ‘Rover’ challenge results, and for making a convincing case for more wide- problem as example: the paper “Engineering a Rover Lan- spread adoption of modeling and MDE. More specifically, guage in GEMOC Studio & MontiCore” by Thomas Degueule, 1) there is evidence suggesting that the quality of documenta- Tanja Mayerhofer and Andreas Wortmann compared language tion of many MDE tools is too low [5], design using a ‘modelware’ tool (GEMOC) and a ‘grammar- 2) while efforts have been made to compare modeling ap- ware’ tool (MontiCore). The paper “A Scenario-based MDE proaches (in, e.g., the Comparing Modeling Approaches Work- process for Developing Reactive Systems: A Cleaning Robot shop [6]), there is insufficient support for evaluating and Example” by Joel Greenyer, Daniel Gritzner, Jianwei Shi and comparing MDE tools, their suitability for specific tasks, and Eric Wete illustrated the use and utility of scenario-based opportunities for interoperation and reuse, and modeling. 3) few repeatable tool evaluations and comparisons exist that Another pair of papers was devoted to identifying promising use appropriate, publicly accessible use cases and that have new research problems. The paper “Modelling as a Service: been carried out by independent third parties. A Survey of Existing Tools” by Saheed Popoola, Jeffrey II. C HALLENGE PROBLEMS AND VIDEO TUTORIALS Carver and Jeff Gray presented a first classification of web- based modeling tools and the paper “Challenges and Research To facilitate the comparison of tools, two challenge prob- Directions for Successfully Applying MDE Tools in Practice” lems were defined by the organizers, called ‘Rover’ and ‘The by Francis Bordeleau, Grischa Liebel, Alexander Raschke, Intelligent House’. Descriptions were made available online1 . Gerald Stieglbauer and Matthias Tichy focussed on how best 1 http://mase.cs.queensu.ca/mdetools to integrate MDE tools into industrial practice. Also, two papers from the Workshop on Human Factors allows, e.g., effective assessment of the product’s current and in Modeling were presented: “Investigating the Effects of future performance and expected lifetime (e.g., preventative Integrating Handcrafted Code in Model-Driven Engineering” maintenance) as well as optimization and improvements in by Tim Bolender, Bernhard Rumpe and Andreas Wortmann, product design and operating conditions. The digital twin is and “Visual Variables in UML: a First Empirical Assessment” thus relevant not only for defect prediction and avoidance, but by Yosser El Ahmar, Xavier Le Pallec and Sébastien Gérard. also to systems engineering and lifecycle management. The Both of these papers emphasized the need for more empirical digital thread, however, refers to the ability to integrate rele- studies, albeit on different topics. vant information from different, traditionally disjoint sources Finally, demos of a tool for scenario modeling (Scenari- to allow for the “right information to be available in the right oTools) and an Eclipse plugin for collaborative modeling using place at the right time” [12]. Realization and use of both the ReMoDD model repository were given. concepts could benefit from the expertise in the modeling community on, e.g., domain-specific modeling; monitoring, IV. D ISCUSSION AND RESULTS animation, simulation; ‘models at runtime’; and support for Overall, the different parts of the workshop were attended different views. However, they also require solutions to the by about 30 people, Most had academic affiliations, but some model integration problem already mentioned above. representatives from industry were present as well. Discussion 4) Human factors: The presentations on human factors was lively, continued past the scheduled end of the workshop, were very well received and triggered comments about the and focussed mostly on the following topics: need to improve the usability of MDE tools and the current 1) Challenge problems: Two challenge problems had been scarcity of empirical studies investigating usability or the defined by the workshop organizers to facilitate comparison impact of modeling. of tools. The problems were phrased in very general terms to allow for the participation of a broad set of MDE tools V. C ONCLUSION that leverage models for different purposes. The keynote had The workshop provided a forum for the exchange of ideas encouraged the use of challenges and contests and shown and identification of challenges and opportunities related to several successful examples. In the 1995, the ‘Production Cell’ the development and use of MDE tools. The use of challenge problem had allowed the comparison of different formal meth- problems was recommended, together with an increased focus ods [8]. On the other hand, the effort required to define suitable on systems engineering, interoperability and integration (e.g., problems was acknowledged. Overall, the use of challenge in the context of OSLC and the digital twin and digital thread problems was considered worthwhile and the formulation of, concepts), usability and empirical studies. e.g., a ‘rover contest’, in which MDE tools are to be used to R EFERENCES develop software for a rover to accomplish some task, was [1] Language Workbench Challenge Workshop. SPLASH 2016. November suggested for next year. 1, 2016. http://2016.splashcon.org/track/lwc2016 2) Systems engineering and integration: The industry par- [2] Transformation Tool Contest. STAF 2017. July 21, 2017. ticipants noted that modeling is extensively used in systems http://www.transformation-tool-contest.eu [3] SAT Competition. 19th International Conference on Theory and Applica- engineering, where it supports the development of the entire tions of Satisfiability Testing. July 2016. http://www.satcompetition.org product and not just its software. Due to the many different [4] VerifyThis Verification Competition. ETAPS 2016. April 2, 2016. kinds of models and tools used, integration of the relevant http://etaps2016.verifythis.org [5] N. Kahani, M. Bagherzadeh, J. Dingel, J.R. Cordy. The problems information contained in different models constitutes a major with Eclipse modeling tools: a topic analysis of Eclipse forums. 19th challenge. It was noted that the ‘Open Services for Lifecycle International Conference on Model Driven Engineering Languages and Collaboration (OSLC)’ effort aims at facilitating this kind Systems (MODELS16). 2016. [6] Workshop on Comparing Modeling Approaches. 16th International Con- of large-scale, product-wide integration and helping organi- ference on Model Driven Engineering Languages and Systems (MOD- zations transition from ‘document-centric’ to ‘model-centric’ ELS13). October 1, 2013. http://cserg0.site.uottawa.ca/cma2013models production processes that are organized around linked data [7] T. Futagami, T. Shimizu, M. Hoshi, J. Tanahashi, Y. Kobayashi, N. Watanabe, T. Yukawa H. Watanabe, Y. Watanabe, H. Makino. ET and model repositories [9]. Despite OSLC’s use of open Robocon: A Software Design Robot Contest for Educating Embedded standards and technology such as the Resource Description Systems Engineers. TNI Journal of Engineering and Technology (2):2. Framework (RDF), linked data, representational state transfer July - December 2014. [8] C. Lewerentz, Th. Lindner. Formal Development of Reactive Systems: (REST), and HTTP and open source tools such as Eclipse Case Study Production Cell. LNCS 891. Springer. 1995 Lyo and OSLC4Net, there appears to be little interest and use [9] Open Services for Lifecycle Collaboration (OSLC). Community web in academic research and teaching. As an interesting aside, portal. http://open-services.net. 2017. [10] J. Bezivin, R.M. Soley, A. Vallecillo. Summary of First International efforts to address interoperability challenges using modeling Workshop on Model Driven Interoperability (MDI’10). Oslo, Norway. were made in, e.g., [10], [11]. October 5, 2010. 3) Digital twin and digital thread: The concepts of ‘Digital [11] P. Grace, B. Pickering, M. Surridge. Model-driven interoperability: engineering heterogeneous IoT systems. Annals of Telecommunications Twin’ and ‘Digital Thread’ were first introduced by the mili- 71(3-4):141-150. Springer. April 2016. tary aircraft industry, but are gaining interest in other domains [12] C. Leiva. Demystifying the Digital Thread and Digital Twin Concepts. such as digital and smart manufacturing and ‘Industrie 4.0’. In Industry Week. August 1, 2016. short, the digital twin refers to a digital model of a product that