Which of these 3 start-ups will crack the complexity problem?

It has now become a mantra: The complexity of development is increasing. Why? Software, networking and compliance. But how can we solve the problem? Agility bought us some time in the 2000s with short development cycles, but did not solve the complexity problem in the long term. MBSE promises a more sustainable solution, but the hurdles are great. Worse, mistakes in the introduction of MBSE can be expensive. That's why I don't think MBSE will become mainstream in this decade. At least not in the way it is being practiced at the moment.

But many start-ups are working at full speed on the complexity problem. Here are a few start-ups that I have been following more closely for a while.

First things first: I have no agreements with these companies, but I know all the founders. If you want to contact them, please tell them that you found them through me.

For each of the tools, I have included the shortest possible manufacturer video showing the aspects described.

Flow engineering: developing the digital twin

Flow Engineering is a UK-based company that currently focuses primarily on the UK market. Flow has a strong focus on requirements and traceability. At first glance, Flow appears to be a modern requirements management tool with a focus on traceability, ease of use and collaboration.

However, Flow also supports its own system view, which exists orthogonally to traceability (In the video from 1:40). Flow can also handle parameters in a similar way to Valispace (in the video from 2:00). However, these parameters come from external models. In the video, the maximum weight requirement is compared with the CAD model, which provides us with the exact weight of the system.

Flow integrates with Matlab/Simulink, Simscale, Onshape, Jira, Excel and a few more and shows how key figures propagate through a chain of models.

Today, we do not yet have an all-encompassing system model, but rather distributed models with mutual dependencies. Flow recognizes the dependencies between models and resolves them. This prevents Flow from working with outdated or inconsistent data (in the video from 3:12). The models can be adjusted directly from the tool and the effects of the changes can be tested.

Spread AI: pragmatically mitigating implementation problems

The Berlin-based Spread AI has been on the road a little longer and recently Investment round of $16 million completed. Spread is particularly widespread in the automotive industry.

Spread has two premises: First, modern systems require three orthogonal disciplines: mechanics, electronics and software. Secondly, development information is located in silos. Furthermore, development is fragmented and scattered across many storage locations.

The solution (from 0:40 in the video) is an "Engineering Intelligence Network" that networks the relevant information, i.e. also traceability. Spread connects to the various data sources and uses them to fill a graph database.

The real highlight is that a low-code module can be used to develop applications that are based on this graph database (in the video from 1:37).

Spread has grown up with the optimization of cable harnesses, so it started way down in the V-model. In this respect, it is not (yet) a solution for system development. Whether Spread can also establish itself at the upper end of the V-model remains to be seen.

Spicy SE: MBSE, but simple and effective

Spicy SE is betting that MBSE is ultimately the solution to the complexity problem. The Intro video from Spicy only describes the value proposition in 60 seconds. It is certainly more interesting for readers to see the tool in action. I have therefore embedded the video below, which shows the work on the architecture. On the website are ten 30-second videos that show different use cases in development.

The premise of Spicy is that classic system modeling, as practiced with Rhapsody, Enterprise Architect, Cameo or Capella, is too complicated. Spicy solves this in two ways. Firstly, it provides a modern, web-based development environment that is designed for collaboration. Secondly, Spicy uses its own modeling language, which is much simpler than SysML.

The second in particular is a bet where it is not yet clear how it will end: is the trade-off in the modeling language between simplicity and power the right one? Of course, I had to think of Capella which also uses a simplified modeling language. However, Spicy's language is (still) a good deal simpler.

Conclusion

Of course, these three companies are not the only ones trying to crack the complexity problem. I'm following a few others, which I may report on at another time.

Dear readers: Which other start-up tools do you know? Please write in the comments!

The exciting thing about these three tools is that they have very different approaches:

• Flow Engineering: Link specific models to an overall system model
• Spread AI: Solving real problems bottom-up
• Spicy SE: MBSE, but simple and effective

What do you think? Will one of these companies clear the field, and if so, which one? Write it in the comments!

Photo by Kenny Eliason on Unsplash