Software has long ceased to be an afterthought in the development of mechatronic systems. At Demcon high-tech systems, software engineers contribute ideas regarding architecture, functionality, and system behavior right from the first customer meetings. According to Laura Claas, Group Leader Software at Demcon high-tech systems, software has grown into a fully-fledged discipline and one of the fundamental building blocks within the multidisciplinary project teams.
Nothing happens without software
Demcon is well-known as a developer of complex machines, mechatronic systems, and high-tech equipment. But behind virtually every machine the company develops, there is also software. Claas, explains: ‘We are a multidisciplinary engineering organization, and software is a crucial part of that. We can create amazing mechanical designs, but if we don’t tell that system what to do, it won’t move. Without software, nothing happens.’
That does not mean that software plays the same role in every project. Sometimes we develop a complete machine including all software; sometimes a customer provides the control platform to which new systems must connect. Even then, software remains an essential part of the project. ‘Software is always an integral part of our projects,’ says Claas. ‘It is just the scope that differs. Some projects are software-light, but never without software.’
software as a building block.
The entanglement between software and hardware has increased significantly over the past decades. Whereas machines used to consist of distinct parts, in modern systems disciplines such as mechanics, electronics, and software are closely intertwined. ‘Twenty years ago, machines were less intelligent, and it was much easier to separate the different domains, allowing developers to work virtually independently of one another,’ Claas reflects. ‘That time is behind us. The best solutions emerge when we develop mechanics, electronics, and software as a single entity from the very beginning. That is our development philosophy.’
Depending on the project, software can play a modest role or be the dominant discipline. ‘Sometimes you have one software engineer in a team with ten mechanical engineers,’ says Claas. ‘But we also have projects where software makes up the bulk of the engineering work, with a number of systems engineers involved as well. That ratio varies by project. What doesn’t change is that software is always part of the bigger picture.’
software from the start.
The change lies not only in the amount of software, but also in the timing of software engineers’ involvement in a development process. They are part of the project team from the very beginning, contributing ideas during the quotation phase, even before a project has officially started. ‘In a three-year project, you can’t suddenly think after two years: maybe we should bring in the software team,’ says Claas. ‘Because it is precisely in the early phase that important choices are made regarding architecture, functionality, and how different parts of a system work together.’
"A good software architecture allows you to extend the lifespan of a product."
This working method is now firmly embedded in the organization. Software engineers are involved in customer meetings and during business development processes, even when it is not yet clear what role software will ultimately play. That broader perspective helps to sharpen requirements and identify opportunities sooner. Business developers, system engineers, mechanical engineers, software experts - they all look at a problem through their own lens. It is precisely by bringing these different angles together early on that we get to the root of the problem faster and can advise customers on the best solution.’
from specialist to connector.
That vision is reflected in the profile of the software engineers at Demcon high-tech systems. Programming knowledge remains fundamental, but the ability to look beyond the boundaries of one’s own field is at least as important. ‘Our software engineers work daily with specialists in the fields of mechanics, electronics, optics, cryogenics, physics, and data. That is why we expect them to understand not only the software, but also how a machine behaves, the underlying physical principles, and how the different disciplines influence each other. It is precisely this combination that makes it possible to develop the best solution together.’
According to Claas, the profession is therefore shifting increasingly towards systems engineering. ‘There is a lot of truth in the stereotypical image of the T-shaped professional. That type of specialist is becoming increasingly valuable,’ she says. ‘Of course, you need your own specialization, but in addition, you need to have a sufficient understanding of the disciplines around you. Otherwise, you cannot make that connection.’
She is already seeing this reflected within her team. For instance, there is an engineer working there with a background in both physics and computer science. It is precisely this combination that proves valuable in complex development projects. ‘We are going to need those kinds of profiles more and more. People who have a fundamental understanding of how a system works and can translate that into software.’
first the behavior, then the code.
To keep the increasing complexity manageable, we are always looking at new methods of software development and incorporating them in our way of working. An important trend in this regard is model-based engineering. In this approach, the focus is not on the programming language, but on the behavior of the system. ‘Traditionally, you quickly think: am I going to program this in C++, C#, or another language? We take a step back and look at what the system actually needs to do. What should the behavior of this machine, this software, or this function be?’
This creates a model that is independent of the final technical implementation. This enables functionality to be implemented in different ways over time, including through AI, while simplifying long-term maintenance and extending the system's lifespan.
‘Programming languages become obsolete. Technology changes. But if the underlying model remains up-to-date, you can implement adjustments much more easily without having to start from scratch every time. Even when a machine has been running at a customer’s site for years.’ You could even use models to generate code through Agentic AI, increasing development efficiency while making systems easier to maintain and keep up to date.
This approach aligns with a broader development within Demcon: model-based systems engineering. This involves looking not only at the software, but at the entire system. ‘In many projects, you record the same information in all sorts of places,’ notes Claas. ‘In the software code, in the electronics, in documentation, and in test plans. You are essentially repeating the same truth over and over.’ By working from one central model, a single source of truth is created. A change then only needs to be implemented in one place and is automatically translated to other parts of the project.
managing complexity.
The model-based approach requires a different structure for the development process. Within Demcon, tooling plays a pivotal role in capturing requirements, planning work, and ensuring quality. ‘For example, we work extensively with Jama, a requirements management tool in which we not only capture requirements but also directly link them to the tests,’ explains Claas. ‘This way, we cover the entire V-model. You create traceability: you can see exactly whether a requirement has actually been realized and tested. And if a requirement changes during or even after the project, for instance because a component reaches end of life, you can immediately identify which components are affected.’
Daily software development is structured accordingly. Requirements from Jama are linked to work items in the project management tool Jira, where teams plan and track their activities. This creates a single cohesive system in which development, testing, and documentation are interconnected. ‘Software engineers on my team don’t have to manually maintain all kinds of documentation. A lot of information is automatically generated from the systems we are already using. That saves time and reduces the risk of errors.’
According to Claas, the added value lies primarily in the structure that results from this. ‘You ensure that people have the right information at the right time. In Jira, everyone can see when something is ready to be picked up and when it is truly finished. This allows you to maintain an overview, even in large and technically complex projects.’
The future is now and it is ours to build
software evolves.
Next to model-based engineering, artificial intelligence is increasingly becoming an integral part of software development at Demcon. ‘AI is already being used to accelerate development, support code generation, perform error corrections, automate compatibility tests, and further automate repetitive tasks,’ says Claas. ‘For us, AI is not a separate experiment; it is already part of our way of working. We are systematically embedding AI into our development processes and applying insights from Demcon data driven solutions in AI and algorithm development to the software development of complex machines.’
There is little doubt that with AI, software engineers will spend less time writing standard code and can focus more on designing structures, evaluating results, and understanding the systems they are developing for. ‘You will use a software engineer less to write code, and more to make the right choices and verify that the generated code is correct.’
According to Claas, software development is entering a new era.
‘The pace of change is unprecedented. AI is already transforming how we develop software today, and its impact will only grow in the coming years. By combining domain knowledge, model-based engineering, and AI, our engineers are fully prepared to stay at the forefront of this evolution.’
