Hi,
For the moment Primtux is an ISO, therefore the proposition to make it something more general.
There’s an Open Source In The European Legislative Landscape and Beyond dev room at FOSDEM 2025. Might be a good opportunity to discuss this further.
Hello,
There is an article from Eset that shows that 32 Millions of computers are still running Windows 10 in Germany. Windows 10 support is ending in October 2025
https://www.eset.com/de/about/presse/pressemitteilungen/pressemitteilungen/security-fiasko-32-millionen-computer-in-deutschland-laufen-noch-mit-windows-10/
It is perhaps the time to invest in some other solution like EU OS ?
Regards,
No, basing off the direct Fedora images is best. It guarantees the fastest updates with no middleman or Github in between.
@rriemann I think basing off HeliumOS could be a better idea. Fedora at scale could be problematic, as there are so frequent updates.
HeliumOS bases off AlmaLinux bootc, which is similar to Fedora Atomic Desktops. The dev is working on a version based on CentOS Stream 10 afaik, not sure about the project status.
These have a way longer support period, and CentOS 10 will also ship plasma 6.
Using the official LTS kernel instead of the CentOS or AlmaLinux kernel would also be a good idea for faster updates and a shorter chain of trust.
https://copr.fedorainfracloud.org/coprs/kwizart/kernel-longterm-6.12
https://copr.fedorainfracloud.org/coprs/kwizart/kernel-longterm-6.6
Maintaining the RPM packaging for such things would need to be done in-house I assume.
Also have a look at secureblue for future security hardening steps. A lot in there can be considered essential.
Secureblue looks promising. Especially as I also thought that BlueBuild may be a good venue and they use the same. So one could just check and import their recipes one-by-one. It also has quite some github stars…
Yes but for professional appliances I would really base off CentOS Stream or Alma, as they are way more stable.
Fedora is not a replacement for Windows 10, Enterprise Linux is.
For a POC I would base off Secureblue, maybe? They remove RPM Firefox which is the single biggest issue with the project. Same with all uBlue Variants that have a name. And you should not use Flatpak Firefox.
It is difficult in the end. I would base off HeliumOS for more stability, and add a few secureblue things on top.
Or do it yourself, copy their build system and directly base off CentOS-Stream, Almalinux or Fedra. For shortest chain of trust and fastest updates.
For the “Enterprise Linux” side of this, it may be worth engaging with the CentOS Hyperscale SIG, who produces a variant of CentOS Stream with KDE Plasma integrated together. They may be able to help on that front.
This sounds like a good idea.
There are some important points from an IT-admin POV like
As important is the perspective of the end user as mentioned before:
Unless the whole organization switches to a FLOSS stack, there will be a mixed environment and this requires interaction between the FLOSS world and stuff like Office, Exchange and corporate printers.
Free software is usually not very compatible with this proprietary stack.
For the end user, in this scenario the people in administration and education, a mixed environment can result in significant friction for daily work tasks.
As someone interacting with public administration IT, I can easily list half a dozen bugs, some related to KDE, some related to other parts of the “linux stack”, that regularly add friction.
These would be the long tail and non-critical bugs, that are sometimes reported years ago, but only affect a very small part of the user base.
Nonetheless these bugs add up if you work in such an environment.
It could be useful to collect and track some of these long tail bugs, maybe even bugs in non-KDE projects, especially from linux users inside the targeted institutions. Next step would be fixing them of course…
Because as soon as you roll-out this EU OS, these bugs will bite and you’ll get a lot of pushback from the affected end-users.
If we can identify bugs which are particularly affecting public administration, it should be possible to apply for STF money to fix them.
Examples for the kind of bugs that are/would be affecting larger institutions are Bug 494981 or this whole category: Akonadi EWS Resource Bugs. Last time I tried, I couldn’t get the KMail/Korganizer suite to work properly with Office. Maybe the latest version is working better. For now, Thunderbird is kind of working.
But there are also many bugs and incompatibilities for example in LibreOffice. Editing PDFs, particularly PDF forms is also quite an endeavour.
Ideally, there would be a cross-project initiative/project to coordinate and tackle such long-tail bugs.
XFA support in Poppler/Okular comes to mind. Do we know how widely XFA format is used within the EU government institutions? I’d say that one might be a showstopper for more broad adoption of Okular, for instance.
They’re pretty common in U.S. government PDF-based tax forms, FWIW.
On this topic, fosdem conf has a talk on Sunday from kde: