James Towers on MBSE: progress and challenges
MBSE has been the subject of much discussion recently, including on this blog. It is useful for readers to be able to compare different perspectives and opinions. James Towers recently published an interesting analysis of MBSE and SysML, which I summarize and comment on here.
James Towers is an expert in MBSE who works for Dassault, among others, and is therefore closely familiar with the further development of Cameo (MagicDraw). The full report can be found online under the title The MBSE Horizon: Advances and Challenges over the next few years.
18 years of cooperation between OMG and INCOSE
The SysML standard is managed and further developed by the Object Management Group (OMG). The OMG recognized early on that systems engineering expertise needed to be included in order to make the standard relevant in practice. This gave rise to the Systems Engineering Domain Special Interest Group (SE DSIC), which was founded 18 years ago. This group has significantly influenced today's MBSE. In particular, version 1 of SysML was not published until six years after the SE DSIC was founded.
SysML 2
As already Uwe Kaufmann recently explained in an interviewone of the biggest changes in SysML 2 is that it is no longer defined as a profile of UML. James Towers also sees this as an important change and also discusses the history of SysML.
From the outset, however, SysML 2 was also to have a UML profile and a mapping between the new metamodel and the profile. This is pragmatic: tool manufacturers in particular were concerned about having to implement a completely new metamodel. However, this also creates additional complexity. Whether this is good or bad for the future of SysML 2 remains to be seen.
SWOT analysis
But to get back to the actual topic, James presents a SWOT analysis on MBSE. This is based on a survey of several experts on the subject.
StrengthsMBSE is a viable approach to developing complex systems
MBSE with SysML now has a solid foundation consisting of community, publications and tools. More importantly, the industry has recognized MBSE as a viable way to manage complexity. Even if the adaptation of MBSE is slow, it is progressing steadily.
WeaknessesMany organizations are not yet mature enough to benefit from MBSE
It is difficult to deliver results with MBSE in a short space of time and generate significant, measurable added value. For example, users expect to get to grips with variant management from day one - this is illusory. This is also in line with the statement from Andreas Willertwho explained in an interview with SE Trends how he was able to increase productivity by a factor of two to three with modeling - albeit over a period of several years.
The tools are also a problem, as Tim Weilkiens commented - these are still stuck in the 90s. Dave Banham also suggested that an OMG tool certification program could help to ensure interoperability.
OpportunitiesMany engineering disciplines can benefit significantly from MBSE
MBSE can be used not only in the technical processes of systems engineering, but also in specific engineering disciplines such as safety, reliability, etc.
Furthermore, clients are now even demanding the use of MBSE, which can act as a driving force. This applies in particular to public contracts. Here, the legislator has the opportunity to act as a driving force.
RisksFragmentation due to missing or late tool support
As mentioned above, the existing tools are already quite old. SysML is expected to be released in 2021, and tools that support the native metamodel will certainly not be available immediately. For the first few years, there will certainly only be UML profile-based transitional solutions. But in the meantime, alternative solutions such as Capella or Simulink, which could increase fragmentation in the market.
Another danger, also mentioned above, is a lack of maturity in systems engineering. If an MBSE project fails in such a situation, the blame can quickly be placed on MBSE, not on the lack of maturity.
Conclusion
MBSE will not disappear - that much is clear. The question is whether MBSE will eventually become mainstream or remain a niche approach. James Towers argues that more and more projects will simply no longer be feasible without MBSE. The question is whether this will also have an impact on less complex projects.
James hopes that SysML 2 will make system modeling easier and more intuitive for experts, which in turn should have an impact on its dissemination. However, it is unclear how modeling experts and tool manufacturers will ultimately react to SysML 2. SysML 2 is a great unknown.
But language alone does not make an MBSE: Competent users and methods are equally important parts. In this respect, James calls on us to use SysML 2 as an opportunity to take up the challenge and drive MBSE forward.
