Then, you don't need any other device, hence "ultimate".
Convergence, Samsung Dex, lots have tried but nothing mature yet. Well, Dex is mature but closed-source and Samsung-dependent. On the linux no-android smartphone side of things, hardware is too low-cost and the phone aspects of linux too brittle.
Aurora is just a new distro...
If we allow ourselves to dream it's not impossible we'll be able to run Windows games on Android in some future :)
I don't want to toggle options and stuff when I don't want to be distracted or interrupted with calls or messages, I just leave my smartphone in another room. Sure you could attain the same with some keyboard shortcuts but you are pretty sure to sometimes forget to do it while not bringing your phone with you is inconscious and always work.
> Aurora is nothing more than a collection of bash scripts, containerfiles and custom programs stitched together.
That sentence appearing on the same page as "rock-solid" is not very convincing either and does not instil confidence.
edit: I re-read my previous post. It is possible it is not clear, there is a level of ease and stability that comes 'image based' distros. I am putting quotation marks, because I am almost wondering if this is the equivalent of leather warm seats in the car. You don't think you need it or want it, because car warms up just fine so its not needed. And yet.. when you try use it once, you are hooked.
The only thing that will enable people to migrate is third-party app support; no matter how good Linux distros get it's all moot if the software users use every day doesn't support it.
It could be very effective for bringing in those who are not particularly computer literate under the claimed guarantee that a random update is unlikely to break the machine. But you would also need significant financial backing and marketing with strong brand recognition to inspire that kind of confidence.
Fun fact, a bit over a decade ago we were probably the first one ever to publish a distribution to rely on btrfs snapshots per default with the Jolla phone. Sadly that did bite us due to reliability of btrfs at the time, and later phones switched to ext4, but with a stable filesystem it's a nice mechanism for handling updates and factory reset.
2) Including homebrew in a Linux distro is a criminal offense normally punished by public flogging.
Homebrew is great and we will be using it a lot more in the future.
I know there were attempts for this before but they were just banking on the fact that the average dev env is an absurd collection of hacked together tools that is both fragile and nigh impossible to set up for new devs. Now they have a financial angle as well /s
Maybe try research instead of reactionary comments.
"Aurora is a paradigm shift for Linux." "Dream about the stars" "Launch a space rocket" - everything about this, down to the choice of the crudely drawn desktop wallpaper, suggests to me that this was done by very young people. If a few kids want to make themselves a "distro" like this, go for it, just don't advertise it as anything more than a simple pet project, let alone a "paradigm shift".
Besides that, IIRC this is based on Fedora, so it stands on the shoulders of over two decades of work on Fedora.
Surely I'd use Fedora Silverblue if I wanted an immutable Fedora.
So much better approach is to get most of what you prefer by picking right distro. On the other hand ublue makes it very upproachable to make such distros (even yourself). Thats why there are so many of them.
I looked at their website and found no DISA STIG documents. I wonder what jurisdiction they’re planning on launching space rockets from?
I wonder if the unspoken “paradigm” shift is the distribution was vibe coded.
There’s a lot of contradictions on the landing page that would easily be explained by either kids writing it, or someone vibecoding the site.
Such as their claim that updates are a “single iso”, and also their claim about a single App Store, and they then go on to discuss flatpak and homebrew package management.
Or their claim to have redesigned the desktop from the ground up, while boasting they run KDE/Plasma.
And there’s also the claims that it brings something totally new while then going on to describe core Linux features.
Also the scripts running “non intrusively” yet that’s just what you’d expect any seasoned admin to do. This isn’t a headline feature unless you’re new to the game.
Good luck to the guys. I hope they enjoy the exercise. But this is definitely a hobby project cosplaying as a serious distro
> Such as their claim that updates are a “single iso”
Updates literally are a "single image" (didn't see "iso" mentioned). Where is the contradiction?
> and also their claim about a single App Store, and they then go on to discuss flatpak and homebrew package management.
There literally is a single app store. Homebrew is not used to install apps, only for CLI tools. Flatpak is the single app store which users use to install apps (through Bazaar). Where is the contradiction?
> And there’s also the claims that it brings something totally new while then going on to describe core Linux features.
Can you explain what exactly you're referring to?
> Also the scripts running “non intrusively” yet that’s just what you’d expect any seasoned admin to do. This isn’t a headline feature unless you’re new to the game.
This distribution isn't targeted at "seasoned admins", so why wouldn't they mention something relevant to their target group? No contradiction here.
Yeah I was typing from memory on phone. So the citations aren’t going to be verbatim.
> Updates literally are a "single image" (didn't see "iso" mentioned). Where is the contradiction?
Because that’s not how homebrew works. And you can’t have a single image if you’re expecting people to install apps via their multiple different endorsed delivery mechanisms.
> There literally is a single app store. Homebrew is not used to install apps, only for CLI tools. Flatpak is the single app store which users use to install apps (through Bazaar). Where is the contradiction?
Because an App Store is ostensibly just a package manager. I get they’re making a distinction between desktop apps and CLI (homebrew does GUI apps too by the way), but when their emphasis is on “easy” and “one way to do things”, having two different ways to install apps contradicts their mission statement.
If they actually cared about this mission statement AND had half the competence they claim, they’d build a unified UI that supports all use cases rather than expect people to learn those different tools and why it matters that they’re different.
> Can you explain what exactly you're referring to?
“Aurora is a paradigm shift for Linux. To rethink the Linux Desktop experience from the ground up, we built Aurora on new technology and principles.”
Bazaar, Plasma, homebrew, etc. none of this is unique to Thor distribution.
They also boast about being able to rollback updates. That isn’t new to Linux either. Though I’m willing to give them the benefit of the doubt that they’ve created a smoother default experience here.
> This distribution isn't targeted at "seasoned admins", so why wouldn't they mention something relevant to their target group? No contradiction here.
i didn’t say thy are targeting seasoned admins. I said seasoned admins would take for granted that’s how you’d write that code. So wouldn’t even consider it something to announce.
The only reason you’d announce it would be because you hadn’t worked in this space before and feel a sense of achievement doing the bloody obvious. (And to be clear, I have zero issue with people having projects like these to learn new skills)
Also, I clearly didn’t say “literally everything was a contradiction.”
I am interested who you think this is targeting. Because they do specifically say this is for developers (amongst other people). And the reason they give (VSCode) is a pretty noob argument. If you can’t figure out how to install an IDE then you’re clearly tech savvy enough to be a developer.
Homebrew or flatpaks don't pollute the base image
I’m sure there is a reason for their design but the messaging is all over the place. They boast about things that you should expect to happen (like testing packages before releasing - even bleeding edge distros do this) and throw superlatives around with little substance to back them up while quoting pretty run-of-the-mill choices like KDE and VSCode. It leaves an overall impression that the people behind it can’t be taken to seriously.
If that’s unfair then I’m sorry. But it’s their job to convince me that I should trust them with something as important as an OS. It’s not my job to give them the benefit of the doubt.
If that distro is even just half as good as it claims, then they need to seriously redesign the entire landing page to be more focused on what those gains are. And I say this as someone who's ran several open source projects myself and has immense difficulties designing landing pages for them. I know it's a hard thing to get right. In fact I think it's actually harder than creating a new distro.
As the other poster said, Homebrew has nothing to do with this. Please read up on how the technology works before declaring this a contradiction.
> Because an App Store is ostensibly just a package manager. I get they’re making a distinction between desktop apps and CLI (homebrew does GUI apps too by the way), but when their emphasis is on “easy” and “one way to do things”, having two different ways to install apps contradicts their mission statement.
You don't install the same things using Homebrew and Flatpak. You install apps through Flatpak, and non-apps through Homebrew etc. There aren't two ways to install apps.
Are you referring to "casks" when talking about GUI apps through Homebrew? Is that even supported on Linux?
> If they actually cared about this mission statement AND had half the competence they claim, they’d build a unified UI that supports all use cases rather than expect people to learn those different tools and why it matters that they’re different.
No, you're just arbitrarily asking for them to make changes based on your misunderstandings of the use cases of each tool.
> The only reason you’d announce it would be because you hadn’t worked in this space before and feel a sense of achievement doing the bloody obvious. (And to be clear, I have zero issue with people having projects like these to learn new skills)
No, that's not the only reason, but you're looking at the project with an extremely narrow lense while not spending any time actually looking into the technology and project, so I can understand that it's the only reason you see.
> I am interested who you think this is targeting. Because they do specifically say this is for developers (amongst other people). And the reason they give (VSCode) is a pretty noob argument. If you can’t figure out how to install an IDE then you’re clearly tech savvy enough to be a developer.
If you'd spend 5 seconds reading up on the technology, you could easily steelman a better argument.
except from a user perspective there is. You have to first consider what type of app you want, and then search for it using the correct package manager.
As I said, if they had a single UI that managed both flatpak and homebrew, then it would be different. Users shouldn’t need to know which technology was used to download and install a particular package - that's a technical distinction that should be abstracted away by the "App Store".
Now I completely understand why they've taken the approach they have. But they've made a technical decision to fragment the UX while advertising the app store for its simplicity.
> No, you're just arbitrarily asking for them to make changes based on your misunderstandings of the use cases of each tool.
I'm not asking them to make any changes and I definitely do not misunderstand these tools (fun fact: I maintain a few open source projects -- so I'm probably more familiar than most with how brew et al actually work).
I'm simply pointing out how their advertising doesn't gel with the reality of the UX they're providing. It is feedback, not a request nor demand.
But for what it's worth, if they did decide they wanted to look into the possibility or a "single pane of glass" for all app management, then KDE already has a tool that might work here and which already supports pulling from different sources via extensions: Discover (https://apps.kde.org/discover). So it might be worth them taking a look at the viability of use that (again, just feedback, not a request).
> No, that's not the only reason
That’s not a rebuttal. It’s just a contradiction.
> you're looking at the project with an extremely narrow lense
I’m really not. I’m comparing it against my 30 years of professional experience with Linux (and UNIX as a whole) administration and highlighting areas where their docs are coming across as amateurish.
I’m open to being proven there there is more going on than appears, but your replies amount to “you’re wrong” without actually providing any detail why.
I run Linux workstations and because I don't get paid for keeping my workstation up to date, I do look for something that's as low-effort to maintain as possible. So it's quite possible I'm the target audience for Aurora. But the project does such a poor job of explaining why I should use this instead of any of the hundreds of other distros.
This isn't me being narrow-minded because, as I said elsewhere, it's their job to convince me that I can trust them with my hardware and my sensitive data. And their site, in it's current state, doesn't do a good job of that. In it's current state, it feels like it's being managed by people who don't have a whole lot of experience in this field.
But as I also said elsewhere, I know better than most just how hard it is to get a landing page right for a project as complex as an OS. So I'm being critical from a place of empathy rather than dismissiveness.
> If you'd spend 5 seconds reading up on the technology, you could easily steelman a better argument.
I was asking you a question. There’s no need to be confrontational with me.
As others have mentioned, would love a more thorough overview and/or a "who is this for".
necovek•2mo ago
While this is the direction many are going for particular use-cases (IoT in particular), I am very much conflicted.
Yes, inconsistent updates between components have caused a couple of nights of fixing my RPM or DEB based systems in my 27 years of using Linux on desktop (but mostly when I mixed sources of packages).
But at the same time, the modern systems thinking is to decouple things to be able to update and upgrade independently. Think distributed systems like web applications. This needs a change in developing components, but once internalized, both improves and speeds up the delivery.
So with traditional Linux distributions already being a mix (small packaged upgrades, but released as a collection - a "release" or "version" of a distribution), this decidedly moves in the other direction.
How does a security fix get quickly applied here? Can one do kernel livepatching? How do you quickly update a component depended on by everything else?
mindcrash•2mo ago
You might find some extensive answers to your questions in the bootc documentation which is the container runtime running at the core of Aurora and other Universal Blue distributions, like the increasingly popular distribution Bazzite for Linux based gaming.
https://bootc-dev.github.io/bootc/
skydhash•2mo ago
sham1•2mo ago
But honestly, since Emacs is so core to my personal workflow, I think that it's fine to use a system extension for it. Alternatively it could be layered on, which would also of course work. After that, interacting with the containers is of course just using TRAMP to "connect" to them, and that of course works just fine.
[0]: <https://github.com/fedora-sysexts/fedora> & <https://fedora-sysexts.github.io/fedora/>
skydhash•2mo ago
I’ve not encountered OS crashes for a long time, and I’m fairly confident on troubleshooting config issues. Image based OS could be fine for single purpose computing, but I tinker a lot on my PC. Anything that is declarative is usually an hassle.
palata•2mo ago
Exactly this. I think I have spent something like 2 hours fixing such issues in the last 15 years.
I don't get it when people say "at least with X I don't need to reformat and reinstall my whole system every year", or "it keeps breaking". I have used Debian, Arch, Alpine and Gentoo, and I really just don't have problems? Lucky me, I guess.
luckydata•2mo ago
palata•2mo ago
For instance, by installing stuff on the system with "sudo make install" that breaks the expectations of the system package manager, or by modifying config files and then not handling the merge conflict during the update, or stuff like this?
Very, very long ago I remember having to reinstall some nvidia drivers once in a while (but while annoying it took minutes), and I haven't used nvidia since then.
dartharva•2mo ago