Last week the German government published a commissioned study on how it depends on software and services vendors (local copy). The day after the publication, Cathrin Schaer of ZDnet called to ask for my thoughts on the study and digital sovereignty for Germany: Is it even possible? Cathrin’s resulting article picked up some of our discussion, but I wanted to take the time here to elaborate on my thoughts.
The original study is somewhat narrowly focused on software vendors and on-premise products like Microsoft Office. While the dominance of Windows + Office remains a hot-button-issue to many, times are changing, and I would like to comment on digital sovereignty in the age of cloud computing. I would expect, for example, Microsoft to eventually sun-set on-premise Office and more or less force customers to move into the Microsoft (Office) cloud. I also would like to move away from the usual Microsoft bashing, because (among other reasons) it has become a viable open source player.
I believe that digital sovereignty is possible for Germany. There are 3 + 1 components to it: (1) software, (2) data, and (3) computing resources. The additional +1 resource is the ability to execute as created by a functioning educational system. In more detail:
- Computers are run by software. Open source has shown to be an important collaboration method (and legal framework) to create high-quality software that rivals and exceeds proprietary software. It is possible to gain more sovereignty, if people are willing to replace proprietary software with open source software. It may not be easy, because people have to let go of their financial, emotional, and capability investment into the old software.
- Data is becoming more important than software, and if only because software is increasingly becoming freely available as open source software. Therefore, access to data is a crucial element for superior service (through software) as such data powers the algorithms. Data requires a vendor to have its tentacles in as many places as possible to ensure the continued flow of data to its servers.
- To run software on data, you need computing power. This is the easiest of the three to crack, because it is mostly a capital investment (i.e. buying servers). However, operating them at scale is also important and requires an educated workforce.
All three, software that meets user needs, data that makes software deliver superior results, and computers to power it all, are only achievable with capable people. Thus:
- You need a capable workforce to achieve digital sovereignty. This calls for continued investment into STEM (German: MINT) education, but with a deep understanding of its significance to society.
A government like Germany’s should invest in creating a viable ecosystem of open source solutions and developers working on those. As the surge in open source foundations shows, this is a trend that has already started. Also, fortunately, we don’t have to wait for the government to act, because leading companies have long understood how they can use open source strategies to keep competitors at bay.
Still, not all companies understand this, and it would be good for public policy to help build that understanding. It is particularly valuable for Germany’s Mittelstand, which is increasingly seeing its margins eroded by software, as I have explained on this blog many times.
Also, the more open data can be generated, the lower the lock-in to proprietary data and services can be kept (allowing for eventual escape). Here, much more investment and understanding is needed, as the open data situation in Germany can be greatly improved. Many bureaucracies and agencies view having to lay open some of their data as a nuisance. The should rather view this as an opportunity for the economy and society to innovate on the data its tax euros pay for.
Finally, more money needs to flow to the various stages of the German educational system, and necessary reform needs to take hold. As recent investments show, not all is bad, but I wish the current influx of money had come earlier and was less focused on buzzwords and more on fundamentals like data and software engineering.
There is more to be said, for example, how startups are an important part of executing a broad strategy and what needs to be done to support them, but this only shows how everything is intertwined and this post is already too long 😉