The more I learn by using Bluefin, the more I realise “I was doing it wrong”…
My path was:
- Home server for TrueNAS
- Home server apps via jails (BSD back then, more recently k8s on TrueNAS SCALE)
- More home server apps via Docker/Portainer
- Started preferring flatpaks on my desktop which I though of as “docker images for desktop apps”
Then I learn about Silverblue and now I have
- Atomic OS with only 5 extra packages layered (and those are for AD authentication to my Samba-driven home domain)
- Citrix installed in a distrobox for accessing my workplace VDI
- My developement projects at home are now in devcontainers for VS Code
I’m getting the hang of it slowly but steadily…
Is it all ideal? No really. My major gripe is with issues in the base image. For example in this thread I discover a bug in a base package which is now fixed in fedora-testing, but there seems to be no way to “override” an RPM of the base system with a newer version. Actually, this is probably technically possible, I just don’t know how to do it.
The number of layers of abstraction though right now, makes it hard for the average user to understand all this. I admire the effort with FAQs like these, that give people some pointers, but I doubt an average user will ever be able to grasp what is happening.
The truth of the matter though is: they should not have to care. The point is that it should “just work” and for the most part simple users just want a stable OS and nice installers. So I would say atomic OS + flatpaks is all they really need… The base system works/updates and they can install whatever office/browser/etc app they need as a flatpak…