Digital transformation: buzzword or substance?

Digitalization is on everyone's lips, I can hardly hear it myself. But people are talking about it for a reason, because there is of course a reason why the topic is being discussed. I would like to shed some light here on what digital transformation means for systems engineering.

What is digitization anyway?

Digitization means moving away from analogue, to put it bluntly. So no paper, but files. But if it were that simple, we would have long since left digitization behind us, because the typewriter has long since disappeared from most offices and is now only used in niche areas.

However, digitalization is also about making the new digital units available in a fully or partially automated way. Here is an example from my area of work: I have been working with ReqIF, the Requirements Interchange Format. This is a file format for the lossless Exchange of requirements. With ReqIF, the important requirement structures are retained, i.e. hierarchies, links and attributions.

In many cases, ReqIF replaces an exchange via Word, Excel or PDF. Although this is also digital content, it is unstructured. In other words, digitization consists of three stages here:

• Level 1: Paper
• Level 2: Unstructured digital form (Word)
• Level 3: Structured digital form (ReqIF)

The digital transformation began with stage 2, but it is only complete with stage 3.

Digital transformation in which scope?

The next question is where the digital transformation should take place. There are two primary areas:

• In development
• In the product

Both areas are important, but are approached differently. The example of ReqIF comes from development, where a lossless exchange format can be used to develop much more efficiently. At the same time, every company seems to feel obliged to digitize their products in some way. This certainly makes a lot of sense for some products, such as a modern smartphone. For some products it may make sense, but it has to be weighed up against the possible disadvantages, such as with connected cars. Because although the benefits are there, the risk of misuse is high. And with some products, you have to seriously ask yourself whether there is any (customer) benefit at all. Like, for example, a Juicer that only works when connected to the Internet is.

Digital transformation in development

At least for the focus of this blog, the digital transformation in development is much more interesting. After all, systems engineering methods can be used to develop both cars and juicers. What is more interesting for us is what the digital transformation means for development. And this has already been suggested above with the example of ReqIF: The goal should be for the artifacts of development to interact intelligently. To achieve this, these artifacts must first be able to be processed by machines.

Standardized metamodels are indispensable for digital transformation [tweetthis]Standardized metamodels are indispensable for digital transformation[/tweetthis]

And this brings us to a topic that has already been addressed here several times: The development artifacts must not only be available in digital form, they must also be able to be processed. However, this is only possible if the structures are at least reasonably standardized and there is at least a rudimentary understanding of the semantics. Although a lot of research and work is being done in this area, there is still a lot to do. For example, although ReqIF is beginning to establish itself, it is only a selective solution to a very specific problem. And even if the structures are standardized, there are no clear semantics for the elements.

Another example is OSLCa service-based approach for integration in development. Although there is somewhat more semantics here and the area of application is broader, the standard is often regarded as over-engineered and there is comparatively little support or acceptance.

Conclusion

This is just my interpretation of what "digital transformation" can mean. I would therefore urge everyone to clarify this in their own context first.

For systems engineering, my interpretation of the meaning is: to enable the artifacts of development to be intelligently interlinked, and models are the way to achieve this.