SysML v2 Plenum: Putting pressure in the right places
Version 2 of the Systems Modeling Language (SysML v2) is currently generating a lot of activity. The standard was recently adopted and manufacturers are working on tools. Many in the systems engineering community are now asking themselves: Is SysML v2 a game changer? Will everything simply continue as before? Or will the standard even die, as users will stick with v1 and tool manufacturers will only tentatively support the new standard?
Eight experts discussed these questions, and more, at the plenary session on SysML v2 at TdSE. Panelists were (from left to right in the picture) Daniel Siegl, Benjamin Baust, Jesko Lamm, Christoph Fischer, Bernhard Kaiser, Mike Nicolai, Bernd Rumpe and Tim Weilkiens. On the far right is Daria Wilke, who moderated the plenary session.
Introduction by Tim Weilkiens
One of the known drivers of the SysML in Germany is Tim Weilkienswho has also published several books on SysML. Tim gave a brief introduction to the motivation behind SysML v2 to give the audience a common knowledge base. He also pointed out the long process that an international standard entails. The process started back in 2009 with the first Request for Information (RFI). The publication will also not be completed until 2024/2025; at the moment, the Finalization phase.
There are many reasons and drivers for SysML v2, but I have already written a lot about this here and conducted interviews.
The plenum
A slide is worth a thousand words - Here are the experts who were on stage once again. Below is a summary of the contributions of the individual panelists.
Jesko G. Lamm
Jesko G. Lamm works as a systems engineer in medical technology and has been an active member of INCOSE for many years. His focus is on system architecture, where he is co-author of many conference papers and books.
In the discussion, he pointed out that SysML v1 carries a lot of ballast around with it. SysML v2 solves this with two approaches. Firstly, SysML v2 is no longer based on UML. Secondly, many concepts have been standardized, making SysML v2 easier to learn.
Jesko urged the tool manufacturers that the tools must pass on this simplicity.
Mike Nicolai
Mike Nicolai sees himself as an innovator with a broad range of technical and non-technical skills. He is currently leading research, development and product development teams for a new product in the Siemens Simcenter portfolio.
As a representative of a tool manufacturer, Mike appealed to his industry not to mess it up. It is not a given that SysML v2 tools will be better than the previous ones.
Mike is convinced that new tools will be web-based and highly collaborative. The file-based data exchange of models practiced in v1 was pragmatic, but should not be pursued further. The SysML v2 API has a lot of potential here.
Finally, he pointed out the importance of domain-specific languages (DSLs) in practice. The improved possibilities of SysML v2 for DSLs could be an important driver for success.
Tim Weilkiens
Tim Weilkiens is a passionate MBSE'er. He is an MBSE consultant, trainer, author of many books, owner of the publishing house MBSE4U and lecturer of master courses. But more importantly, he is an active member of the Object Management Group. Tim contributed to the original SysML specification and is currently co-chair of the SysML v2 Finalization Task Force.
Tim reminded with renewed urgency the trinity of MBSE: it consists not only of the language (SysML), but also tools and methods. Forgetting this is a recipe for failure.
The trinity of MBSE: language/notation, tool and method.
Tim Weilkiens
In the plenary session, Tim pointed out the things that are still missing. In particular, there is currently no diagram exchange format. This was often a problem with SysML v2. There is also a GfSE working groupbut which lacks resources. He appealed to the audience to get involved if possible.
Bernhard Kaiser
Bernhard Kaiser is a specialist in functional safety and systems engineering, with a focus on the automotive industry. He has experience as a developer of embedded systems, project manager, researcher and safety consultant. He deals with the improvement of safety engineering methods and their integration with model-based development approaches. He is not only involved in SysML v2, but is also a member of the committee for the ISO 21448 standard. As an Ansys employee, he also comes from the corner of tool manufacturers.
For Bernhard, the API in particular is a game changer that enables SysML to act as a control center for development. This will also be part of the product strategy at Ansys. He sees the API as a great opportunity for the powerful but somewhat outdated Scade, which originally dates back to the late 1990s. Ansys plans to base the next major version of Scade as a cloud-based solution directly on SysML v2.
Christoph Fischer
Christoph Fischer has a background in medical technology and has spent a large part of his career in architecture. He combines agility and systems thinking in development, from the initial idea to end-of-life issues and cross-project projects.
Christoph also brought some skepticism into play: The SysML v2 specification, which is over 700 pages long, is not really well suited for users. He raised the important question: How should users learn how to use the language? Training is not the desired approach for all companies. Is on-the-job training possible? He also saw migration as a major problem that has not yet been solved satisfactorily.
His wish list included "Copilot for SysML". I personally found this exciting, as the vision of the product Semiant of my company is heading in precisely this direction. Christoph sees many great opportunities for SysML v2, but they require some "friendly pressure" on the tool vendors. His ideas included cross-domain work for cyber-physical systems, simulation and integration with other tools, including PowerPoint (how else are we supposed to reach the decision-makers?).
Bernhard Rumpe
Bernhard Rumpe is Professor of Computer Science at RWTH Aachen University. He focuses on the use of modeling languages in software development. He is co-founder and Editor-In-Chief of the Springer International Journal on Software and Systems Modeling.
You could tell that Bernhard comes from the software sector because he talked a lot about what we learned there. In particular, the Open Source software community, which has led to today's ecosystem in software development: Productivity there has increased enormously thanks to reusable components. Open libraries in systems engineering could be an accelerator here. However, he doesn't know how we can achieve this either. He would like to do research in this area.
Like Tim, he pointed out the trinity of MBSE and said that without a change in development processes, SysML v2 makes no sense.
Daniel Siegl
Daniel Siegl is responsible for business development at LieberLieber and is an evangelist for MBSE and systems engineering. He was already at SE-Trends, together with Peter Lieber, in conversation.
Daniel is in an interesting role, as his company does not develop a SysML v2 modeling tool, but rather extensions that make a partner tool (Enterprise Arcitect) more effective. These extensions compensate for the weaknesses in the tool. As a result, he has an extremely good understanding of the problems faced by users. He is highly motivated, but as a small tool manufacturer, his company has few options.
Accordingly, he also addressed the users in the audience directly and pointed out that it is the users who are most likely to be able to influence the manufacturers.
Dear users, if you want interoperability, you have to put pressure on us. In three years' time it will be too late, then all you can do is moan.
Daniel Siegl
He also gave very specific advice to users. In particular, he warned against misunderstanding the textual notation. This is also "only" a view of the model, not the model itself. And accordingly, it should not be used for versioning in git in order to avoid conflicts with WinMerge to dissolve. That is not possible!
He also called for preparations for migration to begin now. Inconsistent or invalid modeling will be a major obstacle to a later migration.
Finally, he was very concerned about the validation of the modeling guidelines to ensure SysML v2 conformity. Here, too, there was a call for funding for corresponding work (within the framework of OMG or GfSE). In terms of effort, a low five-digit euro amount would be missing here. So if a company with a strategic interest in SysML v2 would like to contribute money here, please contact Daniel or GfSE.
Benjamin Baust
Benjamin Baust works for Bosch and has a background in vehicle systems engineering. He has worked as a systems and software engineer for several embedded control systems - always with a strong interest in modeling, simulation and development methodology. In his current role, he works on model-based system engineering methods with a focus on digital transformation in systems engineering. Benjamin is therefore a user whose company is large enough to put pressure on tool manufacturers.
For Benjamin, textual notation was an important aspect of SysML v2. Not surprisingly, interest came particularly from the software-related areas. He pointed out that in his environment a lot is done with PlantUML is worked with. PlantUML is also used a lot for generating diagrams.
With regard to the implementation, i.e. the introduction and use of SysML v2, he still sees many question marks. He would like more information from the manufacturers. The question of mixed workflows, i.e. the mixture of graphical and textual modeling, was important to him. The experts confirmed that this (and more, see DSL above) is possible with SysML v2.
Daria Wilke
Finally, a big thank you to Daria Wilke, who led the discussion brilliantly. Daria is deputy chair of the GfSE and a doctoral student at the Fraunhofer IEM in the field of systems engineering.
Conclusion
The plenum was a good mix of tool manufacturers, users, consultants and representatives of the teaching community. The discussion showed that the success of SysML v2 is far from assured. Users in particular are now called upon to put pressure on tool manufacturers in the right places.
Image: GfSE
