It's helpful if you have multiple devices, even if it's just in the family home. Saves you having to manage all of them - you just manage one image and they all pull from that. Or you let someone else, like Universal Blue, manage the image - the effect is the same. Want to add a package to all of your devices? Just do it once. Want to change the firewall rules across them, too? Same again. One image = more time for other stuff, and less time being an unpaid sysadmin at home, everything just works.... And if it doesn't, you can rollback and there's only one image to fix 😂

​@@Happy_Chappy Thanks, that was a good explanation for image based distros. I can see it would be helpful for peeps managing other devices, I know an old retired guy doing it for free for other older folks so it would be nice for peeps like that. No one else in my family uses or wants to use Linux so they don't usually bother me with PC stuff 😸

Counter question, why wouldn't you want your desktop to be reliable?

I need reliability in my desktops, it's not an option for me. There's no reason why my desktop should be less reliable than my server, I treat them the same.

​@@JorgeCastro @owlmostdead9492 I don't understand, I have been using the normal versions of Fedora, openSUSE, and Mint for many years. No reliability issues here. I don't buy the claim that image based OS is more reliable than normal distros for a single desktop user who has common sense enough to spin up a container or VM when they want to do some experimentation without breaking the host. As @Happy_Chappy said I now understand the benefits for multiple deployments which a single person has to manage. Once again an enterprise use case mostly but brought to a very small niche of Linux desktop users who actually manage other Linux desktop user devices. I only know 1 person who does that, retired guy who repurposes older folks machines with Linux for free. It would be useful for peeps like him but I know no other desktop Linux user that would benefit from this. Jorge, you seem like a nice guy and I've been following this channel for some time out of curiosity but it seems like an excessive leak of enterprise first solutions into the desktop space, not that there's anything wrong with it, Linux is free software and you're free to do whatever you want after all, it doesn't make much sense to normal Linux desktop folks is all I'm saying. Or for that matter your own use case from past videos I've seen on your setup, basically Ubuntu and some other dev environments as containers right? You could just do that on a normal distro!

Does this existing bother you so much that you need to leave a comment? It's clearly not for you so why care what other people do with their free time?

"Cloud native" is another way of saying your OS is first built and updated *away* from your machine, then if that build is successful it's sent to your machine either through installation, system updates, or rebasing (switching to another image/flavor, fedora in this case). If it isn't built successfully elsewhere, it doesn't pass at all. Hope that helps!

@@phonewithoutquestion80 Isn't that the same as what happens with Fedora? Someone else builds updates, tests them, and then the organization pushes them to the broader public. If it doesn't work on the test machines, it doesn't get released?

@@michaelpolsinelli2345 Kind of. A more accurate thing is to say this is atomic. That is, you get the whole thing delivered to you as a single deliverable. If it doesn't work for you, you can go back a day, multiple days, what have you until it does. You can take an image and pin it and use that pinned image as a "fallback" plan if you need to. You can also set this system up personally if you want to, so that you're delivering a custom known working image to yourself every day that needs very little maintenance or work. Other distributions are multitudes of varying software that come into your machine, but all have different potential versions based on when and how you upgrade. A given Image of Bluefin is a collection of all of those programs frozen in time. If you want to upgrade something that came from or on the image (like, say, vscode in Bluefin-dx), you literally have to get the whole OS delivered to you again. (Well, technically not the WHOLE Os, there's some magic that lets unchanged parts stay chunked together so you roughly only download the parts that change), but the gist of this is, you CAN'T accidentally break it. And if you do find a way it's broken, you can immediately reboot into the version that worked for you. So like, if I have the latest version of Arch, I may not have software someone else has. If I haven't updated, I may have older versions of software and libraries someone else running with the exact same packages I have has. These versions may operate COMPLETELY differently even though we have , basically, the same installed packages and kernels. But if you and I are both on bluefin-stable, the image that makes up your OS and mine is _EXACTLY_ the same. And if we run a lab, every single image in that lab would be exactly the same. And with the software base being the same comes a lot of advantages. Namely, if there's a bug on one person's machine, there's a bug on all people's machines... BUT.. if you fix that bug on one machine, you potentially can fix it on ALL machines for that image. (Weird edge cases exist, like, Nvidia.) But that's the difference, and it's a game changer. It's like how your Phone OS updates or a Chromebook updates, in a whole package.. and not like Arch or other distros update, with partial and interlocking components, sometimes being missed.

I'll work on video on this!

I think the better term would be container-native... because that's what it is. Containers become the heart & soul of the entire system from the kernel level to userland. The OS is delivered as container images, updated as container images, user installable software is also done with containers and even development happens within containers (user development I mean... the distribution development at a deep enough level HAS to happen outside containers)... The claimed advantages are system resilience to breakage and more compartmentalisation which helps tackle the immense complexity in more scalable ways that as a huge collection of packages in a single universe. Instead stuff has separation boundaries at various levels. This vision is not yet fully realised IMO since complexity in software is an overwhelmingly difficult problem to solve. Containers are just one tool in this... the first baby steps so to speak.

If you're using Bluefin or Aurora, we're using bootc already.

Couldn't have done it without you buddy!

Yes, it's included.

I use the steam flatpak but you can do it via a container if you want. There's no native PWA support as we ship firefox by default but it's pretty straightforward to install Chrome or Edge and use it that way. There's no hyperland in Bluefin it's a gnome experience.