“Simplified desktop”…go to their website, literally no pictures of the desktop interface (at least on mobile).
Everyone who makes software, always follow this formula on your homepage: clearly state what problem exists without your software, show your software, clearly state what problems go away with your software, then describe its actual features
The screenshot you're asking for will be indistinguishable from any GNOME desktop screenshot you find on GNOME's website. Before giving advice it would be worth understanding the subject matter. It's a new OS, not a new desktop shell.
The title statement "works toward a simplified desktop" makes the authors' focus pretty clear, so it's just fair to look for screenshots. It would be disappointing if it was just the same as GNOME.
Desktops shouldn't be simple. They should work reliably and predictably like every good procduct. Everyone wants something simple until they want to do something that's not supported because it's not a simple use case. So unless you are a very simple person who always only does very simple things and lead a uniquely simple life (i envy you), a simple desktop is not for you.
There are two categories of things that should be simple in a desktop:
1. Common tasks. These should be simple because you do them a lot. Browsing the file system, launching applications, using a password manager, sharing data over networks, and making backups - these are all tasks that should be common and should be simple for anyone to do.
2. Important tasks. These should be simple because they are necessary, even if they're not common. Installing software, connecting to new networks, adjusting displays, enabling full-disk encryption, running system updates - these should be simple so that anyone can do them when needed, with as little difficulty or friction as possible. (Admin privileges may come in to play for some of these, of course.)
If any of these tasks are not simple, there are a lot of users who simply won't do them. And that's bad for all of us.
No. Simple means being willing to make hard choices and say “no” when something’s inclusion doesn’t warrant the cognitive overhead it adds. Or effectively designing so that complexity is only exposed if you really need it.
Yeah maybe, but for desktop OSes it usually means "we don't want to do the work to implement that obviously useful feature". Consider for example MacOS removing the mouse acceleration setting. What a brave choice.
> designing so that complexity is only exposed if you really need it.
Yeah this is the right way to do things. But again, often stuff is just ripped out rather than sensibly managed. Another example: most of the useful WiFi settings in Linux are not accessible in Gnome by default. You have to install the third party `nm-connection-editor` tool. Why? All that stuff should be accessible from Settings.
> Consider for example MacOS removing the mouse acceleration setting. What a brave choice.
I haven’t used a Mac with a mouse in ages. I feel the trackpad interface is so much better (and consistent between laptop and desktop) that I think a move away from mice is a deliberate choice by Apple.
For gaming, absolutely. But macOS and Apple devices aren't tailored towards that, and their mice are oddballs anyway: one button, and heavily tailored towards gestures. They're a mouse and touchpad in one. I quit using it when I got RSI. Now I use a vertical mouse (by Logitech). Now that is one of a kind (but it does not do gestures well).
Not just for gaming. Mice are better for everything.
> Now I use a vertical mouse (by Logitech).
Ugh I tried that when I had RSI. Absolutely awful. The fundamental flaw is that you click sideways which always moves the pointer a little unless you strain to avoid it, which kind of defeats the point.
Get a better chair and desk. That solved the RSI for me - no weird ergonomic input devices made any difference.
For me, a couple of years ago a vertical mouse almost instantly solved my RSI which I only once had severe before in 40+ years of computer usage (that was due to physical work, though they attempted to gaslight me into being a heavy computer user). Using a mouse, trackpad or trackpad meant trouble. Slowly, it healed, and I can once again use a trackpad. But nowadays I do not use normal mice anymore for any prolonged tasks. Only my vertical one. I had a different one before the Logitech one. Some cheap ass brand. It worked well for a couple of months of heavy usage and then it had hardware malfunctioning. I went with the Logitech and years later still goes strong, with a couple of weeks of battery life.
You're wrong regarding mice being always superior. Mice have their place, and I am not the only person who has benefited from vertical mice (there is a learning curve, btw). Mice have a severe drawback: they need more physical, flat, clean space than the other pointer solutions. Try a mouse with a cyberdeck and tell me how that worked out. I've done pentesting tasks with a small laptop the size of two mice. I would not be able to do that outside on the go with a mouse. I also had situations where back in the days I had no space for a mouse, so I used my trackpoint.
The DEs it uses are GNOME and KDE, which are as far from "simple" in that sense as one can get on Linux-based systems: the two largest (and perhaps most widely criticized for being buggy and bloated) DEs. While the use of uncommon (non-GNU) userland sounds like a stream of unusual issues to debug, adding up to a strange combination.
Though probably not counting the DEs as an important part of the system (and maybe they have other DEs or WMs in the repositories), one may argue that the non-GNU userland is simpler in a sense. But then again, it is presented here as a simplified desktop, while GNOME-based and KDE-based systems are on the images it provides. And it lists GNOME as its primary DE [1].
> one may argue that the non-GNU userland is simpler in a sense.
This seems to be the main argument of the article. Along with some stuff about systemD.
> In service of that goal, the project is based on BSD tools. Chimera's frequently asked questions page explains that unlike other projects that use those tools for licensing reasons, project picked BSD tools for their smaller code size and reduced complexity.
> But then again, it is presented here as a simplified desktop, while GNOME-based and KDE-based systems are on the images it provides.
Yeah, when I read the title here, I thought it would be about a simplified desktop environment. But they seem to just be using the standard stuff. So, I don’t think this is what they meant to say.
Take Systemd for example. Chimera Linux was to implement the same functionality and be full featured. It wants to do this with a simpler, more modular, more understandable, and more maintainable design.
>They should work reliably and predictably like every good procduct.
So you're saying they should be simple?
>unless you are a very simple person who always only does very simple things and lead a uniquely simple life
Considering the massive popularity of MacOS and iOS which all mandate the same reliable and predictable user experience on every Apple device, most people are "very simple" persons who "do very simple things and lead a uniquely simple life".
The power user paradigm of old that expose all the knobs and dials and levers there are to pleasure us simply does not appeal to the commons.
For iOS there’s absolutely an argument that it’s not capable enough, but as a regular user of all three major desktop OSes I find such complaints about macOS overblown. While there’s small kernels of truth here and there much of it comes down to macOS being built around a different set of conventions than the desktop environment they’re most accustomed to (usually Windows) than inherent incapability. This is further evidenced by how it’s common for longtime Mac users have similar complaints about the Windows desktop being inadequate/incapable.
Desktop environments and user workflows are insanely personal things, not unlike clothing, diet, and music preferences but for some reason many in the tech sphere refuse to acknowledge this and try to position their preferred environment as objectively more correct/superior/etc. It’s really tiring.
> The power user paradigm of old that expose all the knobs and dials and levers there are to pleasure us simply does not appeal to the commons.
I feel that I wasted years of my life changing UI themes and colors since Windows 3 and MacOS 7, and, frankly, I have never felt tempted to do anything like that since Gnome Desktop. On the more vanilla Gnomes I don’t even change the wallpaper.
Except for most ordinary use cases. I guarantee you that, if someone downloads a compressed file, they'll find it infinitely easier to right click and select the "Extract here" option than whipping out some tar command. The linux-brain has people convinced that the layman would prefer the latter if only they would listen.
That's some windows-land programming right there. The idea that a file is something clickable, some icon with a title somewhere on the screen.
And typing 5 letters is clearly not infinitely harder than right-click and rummaging through 20 item menu to find the right action. The difference is actually infinitesimally small.
I'll hold your hand while I offer this groundbreaking revelation: the vast, vast majority of people are in Windows-land. We're even having issues with the younger generations who do everything on their phone and this don't really have any concept of file systems. Get out of your echo chamber. That means getting off of HN and start looking at how laymen use computers.
One of the side benefits of using the FreeBSD userland is all the flags will be consistent regardless of the of program. ‘-l’ or ‘-a’ are going to do what you expect irrespective of the command. It’s nice, and I used free for along while. I still know BSD syntax better than the SystemV equivalent.
I have also used the GNU tools but the BSD ones are winning me over. Another benefit is consistency with macOS I suppose. It also uses the BSD userland.
Simplicity in implementation is not necessarily the same as simplicity in usability. Both are desirable, but if there's a trade-off between the two, then I reckon it's quite justifiable to prioritise the former.
Try installing OpenBSD then. Even though the installer runs in text mode, it holds your hand thru a standard setup, all the way to a graphical login. The basic X11 desktop is a part of the base system (as opposed to packages/ports).
The installer also supports a bunch of advanced features, like scripting an automated install. OpenBSD is also some of the cleanest and simplest code I've ever read.
Maybe it seems a stupid question, but it's an honest one: why openbsd rather than freebsd?
I tried freebsd when Debian integrated systemd. Being neither savvy or intelligent, I had trouble with the ports system and installations were taking hours for small applications. I gave up.
I've always had a nagging desire to try again, but poor health has discouraged me from those kinds of learning curves. I recently left Debian again and use Void, with Opensuse on a separate drive for purposes of maintaining some competency and familiarity with systemd (just in case...).
Anyway, I'd appreciate your input on the various BSDs. As I understand, recent funding suggests it may be growing and will remain viable for a long time, which was, however relevant or not, a concern for me.
Edit: I might also mention that I gave serious consideration to Chimera before settling on Void, for similar reasons of learning curves. Of all distros, I had determined it would either be Void or Chimera and nothing else - or BSD.
> I recently left Debian again and use Void, with Opensuse on a separate drive [...], I gave serious consideration to Chimera before settling on Void [...]
Have you tried to write down your goals/expectations? It seems like you're throwing distros at the wall to just see what sticks. I don't care for apt vs xbps or systemd vs openrc. My own goals are simple:
- I want a "UNIX workstation", overall decent for software development, need to run complex proprietary software (Logic, Compressor, Photoshop...), interface with niche hardware, develop for iOS, and occasionally play games: macOS
- I want to play a couple of games that won't run on Mac/Linux/BSD: Windows
- I want a simple system that I can understand and actually enjoy using: OpenBSD, Alpine
> purposes of maintaining some competency and familiarity with systemd
You can also just use a VM or a homelab box. No need to care for your personal device, as long as it fulfils your stated goals.
> why openbsd rather than freebsd?
I think OpenBSD is to FreeBSD what FreeBSD is to Linux, and then some more. I'd say these are its most distinguishing features (compared to almost every other OS):
- User-friendly - assuming you have some experience with the command line, can follow prompts, read documentation, etc.
- Excellent documentation, check out the FAQ: <https://www.openbsd.org/faq/>; manpages: <https://man.openbsd.org>; etc. Even compared even to the FreeBSD handbook, which (while a step up from most Linux distros) isn't quite on the same level.
- Secure by default: it is the upstream project for OpenSSH, LibreSSL, PF (notably forked into macOS and FreeBSD), and some others; also: privilege separation, W^X memory, (K)ASLR, doas, signify, pledge, unveil, kernel/libc random relinking, syscall origin verification, etc. Best part, you get all of that for free and you don't even notice. There's no "hardening guide".
- Truly self-hosting - the base system has everything you'd require to comfortably develop the base system, including X11, clang, mg (an Emacs clone - if you're not a vi fan), tmux, etc. You will probably install packages anyway, to get your preferred tools, but the base system will still be there.
- Innovative, not just technologically but culturally: cvsweb was the first of its kind - the source code, including all commits, could be anonymously browsed via the web, in (IIRC) 1996; most contemporary projects would just occasionally throw tarballs over the fence.
- Simplicity. It is not just the manpages or the source code; OpenBSD meticulously rips out or refuses to accept complex systems (sudo, gpg, seccomp) and instead introduces simpler interfaces (doas, signify, pledge)
I wish I could live in a world where this is all I need, but unfortunately, OpenBSD brands itself as a "research OS", and doesn't provide any accommodations such as binary compatibility between releases, Linux syscall emulation (which FreeBSD does), and so on. Even vmd(8) is a recent addition.
Hell of a reply. I'd definitely buy you a coffee for it.
Thanks for all the considerations.
PS: regarding throwing distros at the wall; I used the same installation of Debian for 10 years. It was my workstation and everything else. But its direction isn't mine anymore, and I gave quite a bit of thought to my choice, ie Void. FreeBSD would be in addition, not a replacement. Also, my primary needs are far more primitive than yours or what you mentioned. I mostly need something I can trust, that doesn't change for the sake of change. That and for it to be able to do basics, eg python without dependency hell, video editing, office, music/editing, caveman AI, and such. Hardly more. Oh, and freedom. That's really important.
I've also been on Debian for 15 years, before switching to macOS. I had a dozen very good reasons (incl. I've finally been doing the work I've always wanted to), but the "change for the sake of change" seems kinda Apple's motto nowadays, and yes it's hella irritating. Still - for me, macOS beats any other OS in that it fits most of my needs almost exactly.
The only true freedom is to do what you desire. For me, it's e.g. playing StarCraft. Yeah I can make it run on Linux under Wine. No, figuring out why it's broken again is not my idea of free time ;)
You might like Chimera Linux. It also “refuses to accept complex systems”, except it is still Linux and so does not have the limitations of OpenBSD.
Chimera Linux also uses doas. The Turnstile project is a good example of Chimera Linux rolling its own when no suitable alternative is available (to replace Systemd in this case).
I see your points, but Alpine already checks most of these boxes, so for me the cost of switching outweighs any expected benefits. (I don't believe in "just trying it out" in a VM, because the real issues crop up only when daily driving.)
I assumed this was a post about ChimeraOS, a gaming focused Linux distro, but it seems this is talking about a totally different Linux distribution with the same name.
“The system also has no relation to ChimeraOS, besides the unfortunate name similarity. ChimeraOS used to be called GamerOS and renamed itself to ChimeraOS later; however, at this point Chimera Linux was already in public development with its name in place.”
It's the userland that always sends me back to Linux. I've tried BSD a few times, but there are enough gnu flags in my muscle memory for command line utilities that it's painful.
This is an honest question. Do you have any examples?
I have always been a GNU user but have been using Chimera Linux. I ran into differences with ‘sed’ and read about a regex difference with “find’. Really curious what flags other people are using.
one sticking point for GNU users can be that BSD utils are more strict about positional order of e.g filename parameter and input options.
rm ~/foo -rf || rm -rf ~/foo
I was forced to get used to BSD syntax when I switched to MacOS but now I prefer it...
Even as a GNU user, I never would never have thought to put filenames anywhere but the end. I mean, with wildcards, it could be any number of filenames.
You are right though, that is an important difference.
It’s unnatural to put the flags after the command. But sometimes I remember I’ve to use a flag after I’ve entered the input; other times, I notice a flag’s omission after executing it, and Control + P puts my cursor at the end, after the input.
A pleasant surprise - the KDE version works "out of the box" on my 2015 MacBook 12" (A1534). This is an unusual machine, and Linux distros tend not to support its hardware out of the box. Chimera even sees the internal SSD, traditionally a stumbling point to installing Linux on this machine.
It has been amazing to read in this thread how many people seem to think that “nothing new in the UI” equals “nothing interesting about this distro”.
It really highlights how the relationship with computers is changing.
The goals of Chimera Linux have little at all to do with the UI. In my view, Chimera Linux is highly innovative and disruptive. Yet, some people can look at it and genuinely wonder what it does differently or what the point of it is at all. Fascinating.
> Chimera supports many different configurations, leaving the user free to carve up their disk as they please — but does not support having /usr on a separate partition.
On systems with both a hard drive and ssd (since on most distros /(s)bin and /lib* are sym links into /usr, most binaries are ultimately in /usr), my recent preference is to have the hard drive as / and ssd as /usr (then sym link or bind mount there as needed for things that could use the speed, like Steam). Am I the only one who thinks this way or am I way off?
I don’t think of ssd and hdd for different parts of the filesystem. Rather / is a zfs/btrfs/bcachefs pool, and the ssd is added as a read cache to the pool
Edit: it can also be a write cache but that’s more tricky, usually with a battery backed hardware raid it’s fine
GNOME? It is super simple. I have used Fedora/GNOME on my elderly family members laptops for many years now, and they just get it. Even the ones who came from Windows 10. The hours I spend on support has dropped significantly. Windows is such a hassel, and its desktop design philosophy is just not that great for casual users.
IMHO, GNOME is not simple, it's dumbed down. They've obviously tried to copy Apple where it suited them, but missed on a whole bunch of important details, like a unified menu system, a powerful terminal emulator, desktop icons (omg), while badly aping all of the worst parts - from requiring 4 clicks to shutdown/reboot (where macOS requires 2), thru a mostly useless top bar that steals the real estate from browser tabs (Fitt's law), to asking SDL to link against libadwaita to draw window decorations. And this is the worst part, they not only do not want to accommodate their users, but also ignore the developers - those who wish to integrate with and therefore empower the free desktop ecosystem.
Apple can get away with all of that because they're a trillion dollar company, but unlike Apple, the power of the open source community doesn't stem from an unimaginable pile of cash, but from interoperability and cooperation.
I tried to use GNOME for about 8 months but there was just too many WTF moments (one caused by Ubuntu's own dock extension). There's plenty to love about GNOME but missing features, bugs, and design/usability issues makes it feel like beta quality software.
A few months ago, someone wrote a blog post[1] cataloguing many of these issues. One thing not mentioned was the lack of a caps/num lock on-screen indicator, this is a feature that is present in GNOME 2, MATE, XFCE, Cinnamon, KDE, and Windows 7/8/10/11 out of the box and toggleable in their settings window. The only way to gain this functionality in GNOME is through a third-party extension. Many laptops continue to be manufactured without a caps/num lock indicator on their keyboard, it's insane this isn't a supported feature in GNOME.
I don't get the feeling that GNOME team has ever implemented accessibility research in their design choices. For those with disabilities, GNOME is unnecessarily difficult to use [2][3].
So far, my experience with people above 60 is that they quickly understand how to find and navigate between the apps they use every day. The ones coming from Windows adapt suprisingly fast, and are very pleased that the fullscreen popups and forced updates are gone. Shortly before I moved them over to Fedora, they started developing a fear of using the computer because they never knew when it would 'lock them out' because the screen was filled with a fullscreen Office365 ad that had no obvious exit button.
The obvious answer is to build on the Linux kernel. That means more extensive desktop hardware support. That means better desktop software compatibility. That means robust support for OCI containers.
But Chimera goes beyond the BSD model in several ways. For example, it aims to bring the Systemd feature-set while avoiding Systemd.
It also uses pipewire and Wayland.
As mentioned elsewhere, the Chimera Linux founder also found the FreeBSD packaging system to be lacking.
Chimera Linux also aims for stateless /etc and /var.
There is a lot more to Chimera Linux than the userland.
Is this a real question? Instead of justifying the choice to use something good without taking the bad bits too, let me pretend you asked “what are the advantages of using the BSD userland”.
I am not a Chimera dev and they have answered this but here is what I understand….
You could argue that the GNU utils are bloated and over engineered. But that is just an opinion. As the Chimera devs point out, there is a “chicken and egg” dependency problem with the GNU utils when trying to build from source and / or in a container. The BSD utils solve this problem.
At the same time, the BSD utils are more complete and powerful than, for example, Busybox or other alternatives to the problem above.
The BSD userland solves one of the problems that Chimera Linux is trying to solve. The BSDs do not solve many of the other problems and so using everything from BSD is not an option.
I will point out that Chimera Linux and FreeBSD both use Clang/LLVM as the system compiler. Chimera uses it for LTO and certain security features.
Definitely going to take this for a spin. BSD has been appealing to me for a long time, but the driver situation is wanting. In theory, a more integrated BSD like experience but with Linux kernel/drivers? Interesting!
I appreciate what the Chimera authors are trying to achieve, but I would never consider going back to a distro that doesn't support atomic upgrades and seamless rollbacks, ala NixOS, Guix, etc. Packaging issues and incompatibilities are inevitable and impossible to prevent. Giving users the peace of mind that their system can always be reverted to a known working state is priceless, and something all operating systems should have.
Making the system configuration declarative, reproducible builds, etc., would also be a bonus, but I wouldn't consider those hard requirements.
I realize snapshots are a feature of some filesystems, which partially addresses this, but I would rather have this feature at the OS level.
Fair enough. It is worth talking about how the Chimera Linux package manager works though.
There is a file, /etc/apk/world that lists all the packages you have explicitly installed. When you add and remove packages, all it really does is change this list. Then it runs the solver and installs the package versions and dependencies required. You could move /etc/apk/world to another system and it would result in exactly the same set of packages between the two.
Replacing /etc/apk/world to any previous state will “roll back” the system to that point.
The package manger is transactional and apk commands either entirely succeed or fail. When I was testing Chimera in a VM, I ran out of drive space during an install of dozens of packages. The system was left in the same state as it was in before I ran apk.
Cports itself is very declarative. Also, unlike most distros (including Arch) there is no separate install and update step. Other than the kernel, and maybe a couple others, nothing runs after the package install.
Finally, Chimera Linux is aiming for stateless /var and /etc.
Chimera Linux may be closer to what you want than you expect.
> Replacing /etc/apk/world to any previous state will “roll back” the system to that point.
That's not quite what rollback means, though.
Say that a new version of a display driver crashes my display server. With NixOS I could reboot into a previous configuration and have a working system in no time. If I have to keep track of /etc/apk/world changes myself, and boot into recovery mode or chroot to fix this manually, it takes time, effort and frustration I'd rather not have to deal with.
Stateless /var is a good idea, but I'm not sure it would address this problem.
In any case, I'll definitely keep an eye on Chimera. We need more distros that are not derivatives, and try to do something different. Wishing the team the best of luck, and hoping the project succeeds!
It's stance on keeping it's own reimplementations of parts of systemd as a dependency sadly rules it out for me. It would be a great choice for a linux jail on freebsd.
I don't. Google's Android support has just slowly eroded away all the Linux benefits with layers and layers of user hostile "security" without the benefits (i.e. backing up my phone to Google doesn't mean a new phone will one-click recover itself to how it was).
We could have something like android in the linux world if manufacturers did the minimum to support Linux (which already accept binary blobs). But they also want to push their (bloated) apps alongside the driver.
I really wanted to switch to this but unfortunately it doesn't have support for Xwayland scaling :( in case youre using 4k, On Wayland only KDE and Gnome seem to have gotten this working.
If you're on KDE check out Karousel. It's a KDE Kwin script that does this type of scrollable tiling right inside KDE. Works great on my ultrawide monitor.
I think it would be better to say “simplified desktop Linux”. The project is more about the Linux “plumbing” than about opinions about the desktop environment.
Chimera is not even out of beta and it already includes GNOME, KDE, XFCE, LXQT, and others. The founder uses GNOME but Chimera users do not have to.
> Surely, 2025 will be the year of the Linux desktop!
I don't know if that joke was more funny in the late nineties when Linux was basically nowhere (I was already using Desktop Linux back then though) or if it's more funny today, when Linux powers the entire real world.
I mean: Linux powers 500 of the Top 500 supercomputers, billions (if not tens of billions) of devices, rockets (SpaceX uses Linux extensively), servers all around the world, routers in every single household or near that etc.
Basically: Linux powers everything besides the desktop.
Somehow I think there's some irony in thinking that joke is still funny in the same way it was funny in 1998 or so.
It is a reoccurring joke, aka a meme (before internet memes were internet memes). If you liked it, it builds on repetition and nostalgia. If you disliked it, you won't like when repeated. On top of that, whether you liked it or not it may get boring.
Nowadays, it is out of date. Reality surpassed it: 99,9%+ of all smartphones is powered by either Linux or Unix (macOS). Are these desktops, with a pointer device? No. But they are GUI clients with touch UI. The Windows version? Too little, too late.
Same counts for Linux desktop: too little, too late. If it ever wants to overtake Windows or macOS it has to be substantially better. This is why some FOSS people dislike efforts like Cygwin, SFU, and successors (such as WSL). It allows Windows to use unique utilities which were tailored for FOSS OSes.
It's the "...on the desktop", i.e. that the general public would have a stated preference for Linux because it served them better than proprietary OSes, putting their makers out of business, that makes it funny. Linus even said "Really, I'm not out to destroy Microsoft. That will just be a completely unintentional side effect." in a NYT interview back in '03. Much like "640k ought to be enough for anyone." and "There will be a market for only 5 computers in the world." (only in this case Linus is actually on record saying it), it's a testament to the overweening hubris of the early FOSS movement.
The things that you list matter only to a small number of nerds and the general public wouldn't care a bit if they were all replaced by FreeBSD, Fuchsia, or TempleOS tomorrow. Even for the consumer applications, Linux is being primarily chosen just because the hardware manufacturer didn't want to pay money to license another OS to benefit their profit margins as much as anything else.
So, yeah, the joke is still funny and it looks like it will be funny for the next few decades too.
Void Linux is part of the status quo they were unhappy with.
Chimera Linux started off as a rewrite of the Void source packaging system. One of the complaints was using shell scripts for package templates. Chimera cports uses Python. Cports builds the binary packages (for apk 3).
TBH depending on Python also isn't the best choice - it has become a huge and complex language, goes as far as optionally embedding another full language (Tcl, for Tkinter), all the async stuff, kitchen sink standard library, etc. I don't even know how much effort it would require to bootstrap.
I would personally go with something like Lua - small, trivial to embed, multiple implementations to pick from, etc. It would fit exactly where the shell falls short.
it's absolutely the kitchen sink standard library and "complex" language that makes it worth using, because 1) it means no additional dependencies and 2) the language is expressive enough to let the template syntax remain simple and well-abstracted
have you ever tried bootstrapping python? because it's really not a big deal (the only hard dependencies are a C compiler, libffi, and zlib, though distros also add libedit or readline, expat, sqlite, bzip2, xz, and openssl, all trivial dependencies present in more or less any system); tkinter is pretty much always compiled and packaged separately
meanwhile with lua you'd need many additional modules, and the templates would be verbose and far less readable
a non-exhaustive list of stuff python provides that would need to be provided separately:
Don't get me wrong, I love Python, and I've been a user since the 2.4 days. You brought up plenty of good reasons to use it - and my own rule of thumb is to rewrite a shell script in Python once it goes past 100 lines.
My point is that Lua-the-language provides similar expressive power to Python-the-language (including basic stuff like lookup tables), while remaining smaller and simpler to understand than the POSIX shell. It's also simpler to vendor than Python, and you probably want to vendor this kind of stuff to avoid the pain of bootstrapping / circular dependencies.
Now I haven't built a build/packaging system like xbps or apk (few people did), but I've been in charge of devops/releng at work for a decade - and if I learned anything at all, it's that it's better to start with the simplest tool that can do the job, and slowly add what you need. E.g. I wrote a tool in 2016 to replace Ansible in my team, and all of this time it's just been doing its job: <https://github.com/rollcat/judo>. Turned out nobody actually needed the kitchen sink.
note: i've been doing lua for two decades and am well familiar with the language itself as well as its inner workings (far less so with python), and i've been building distro tooling for almost as long, and no, definitely not
I do enjoy Lua, what I miss the most is a more complete and direct interface to POSIX - possibly as a compile-time option, to accommodate the existing embedded use cases.
sure, in fact i updated an entire distro to a new python major version several times
cbuild started iirc with python 3.8, currently it requires at least 3.12 due to some features (it will stay on that minimum for a while, though newer versions are always supported), updating it has always been seamless though
At the PPC prices, you got to have a really good reason to want a PPC desktop. The TalosII cost nearly $10k for the base 2X 4-core Power9 CPU version, and another $5k to upgrade to the 22-core CPU version.
And this is why, eventually one gets tired of all reboots, and endless variants on Distrowatch, get a computer at BestBuy, Mediamarkt, Dixons, or whatever is the local chain, run GNU/Linux on a VM, and be done with it.
Basically, had it not been for AT&T lawsuit, or Microsoft losing interest with either Xenix or Windows NT/POSIX subsystem, and Linux would hardly matter.
You're pointing out, in other words, that UNIX/POSIX is what matters, regardless of the package.
“Simplified desktop”…go to their website, literally no pictures of the desktop interface (at least on mobile).
Everyone who makes software, always follow this formula on your homepage: clearly state what problem exists without your software, show your software, clearly state what problems go away with your software, then describe its actual features
The screenshot you're asking for will be indistinguishable from any GNOME desktop screenshot you find on GNOME's website. Before giving advice it would be worth understanding the subject matter. It's a new OS, not a new desktop shell.
The title statement "works toward a simplified desktop" makes the authors' focus pretty clear, so it's just fair to look for screenshots. It would be disappointing if it was just the same as GNOME.
Probably meant as "Desktop OS" not "Desktop UI".
A Linux distribution meant for desktop use as opposed to server use.
To quote G.B. Shaw: "The single biggest problem in communication is the illusion that it has taken place"
well that's just what the LWN article states, the project itself does not do that anywhere
the desktops on chimera are standard (gnome/kde/xfce and a variety of compositors and window managers available for installation)
Towards a “simplified desktop Linux distribution”, not a “simplified Desktop Environment”.
The screenshot could be GNOME, KDE, or something else. It would not help you understand what Chimera Linux is trying to achieve.
Apparently the article doesn't help either.
Desktops shouldn't be simple. They should work reliably and predictably like every good procduct. Everyone wants something simple until they want to do something that's not supported because it's not a simple use case. So unless you are a very simple person who always only does very simple things and lead a uniquely simple life (i envy you), a simple desktop is not for you.
There are two categories of things that should be simple in a desktop:
1. Common tasks. These should be simple because you do them a lot. Browsing the file system, launching applications, using a password manager, sharing data over networks, and making backups - these are all tasks that should be common and should be simple for anyone to do.
2. Important tasks. These should be simple because they are necessary, even if they're not common. Installing software, connecting to new networks, adjusting displays, enabling full-disk encryption, running system updates - these should be simple so that anyone can do them when needed, with as little difficulty or friction as possible. (Admin privileges may come in to play for some of these, of course.)
If any of these tasks are not simple, there are a lot of users who simply won't do them. And that's bad for all of us.
No. Simple means being willing to make hard choices and say “no” when something’s inclusion doesn’t warrant the cognitive overhead it adds. Or effectively designing so that complexity is only exposed if you really need it.
Yeah maybe, but for desktop OSes it usually means "we don't want to do the work to implement that obviously useful feature". Consider for example MacOS removing the mouse acceleration setting. What a brave choice.
> designing so that complexity is only exposed if you really need it.
Yeah this is the right way to do things. But again, often stuff is just ripped out rather than sensibly managed. Another example: most of the useful WiFi settings in Linux are not accessible in Gnome by default. You have to install the third party `nm-connection-editor` tool. Why? All that stuff should be accessible from Settings.
> Consider for example MacOS removing the mouse acceleration setting. What a brave choice.
I haven’t used a Mac with a mouse in ages. I feel the trackpad interface is so much better (and consistent between laptop and desktop) that I think a move away from mice is a deliberate choice by Apple.
It's very good but mice are still undeniably better. Even Apple sell mice.
For gaming, absolutely. But macOS and Apple devices aren't tailored towards that, and their mice are oddballs anyway: one button, and heavily tailored towards gestures. They're a mouse and touchpad in one. I quit using it when I got RSI. Now I use a vertical mouse (by Logitech). Now that is one of a kind (but it does not do gestures well).
Not just for gaming. Mice are better for everything.
> Now I use a vertical mouse (by Logitech).
Ugh I tried that when I had RSI. Absolutely awful. The fundamental flaw is that you click sideways which always moves the pointer a little unless you strain to avoid it, which kind of defeats the point.
Get a better chair and desk. That solved the RSI for me - no weird ergonomic input devices made any difference.
For me, a couple of years ago a vertical mouse almost instantly solved my RSI which I only once had severe before in 40+ years of computer usage (that was due to physical work, though they attempted to gaslight me into being a heavy computer user). Using a mouse, trackpad or trackpad meant trouble. Slowly, it healed, and I can once again use a trackpad. But nowadays I do not use normal mice anymore for any prolonged tasks. Only my vertical one. I had a different one before the Logitech one. Some cheap ass brand. It worked well for a couple of months of heavy usage and then it had hardware malfunctioning. I went with the Logitech and years later still goes strong, with a couple of weeks of battery life.
You're wrong regarding mice being always superior. Mice have their place, and I am not the only person who has benefited from vertical mice (there is a learning curve, btw). Mice have a severe drawback: they need more physical, flat, clean space than the other pointer solutions. Try a mouse with a cyberdeck and tell me how that worked out. I've done pentesting tasks with a small laptop the size of two mice. I would not be able to do that outside on the go with a mouse. I also had situations where back in the days I had no space for a mouse, so I used my trackpoint.
did you ever try a TrackPoint, with a HARD inverted dome, and mouse buttons under the space bar? Much faster for me than a mouse any day of the week.
You mean the IBM nub? Yeah I have used them. They're even slower than a touchpad. Probably the slowest kind of mouse there is!
Maybe, but if you touch type and click at the same time, they’re faster.
I think they mean simple in the sense of minimized complexity, rather than dumbed down.
The DEs it uses are GNOME and KDE, which are as far from "simple" in that sense as one can get on Linux-based systems: the two largest (and perhaps most widely criticized for being buggy and bloated) DEs. While the use of uncommon (non-GNU) userland sounds like a stream of unusual issues to debug, adding up to a strange combination.
Though probably not counting the DEs as an important part of the system (and maybe they have other DEs or WMs in the repositories), one may argue that the non-GNU userland is simpler in a sense. But then again, it is presented here as a simplified desktop, while GNOME-based and KDE-based systems are on the images it provides. And it lists GNOME as its primary DE [1].
[1] https://chimera-linux.org/docs/
> one may argue that the non-GNU userland is simpler in a sense.
This seems to be the main argument of the article. Along with some stuff about systemD.
> In service of that goal, the project is based on BSD tools. Chimera's frequently asked questions page explains that unlike other projects that use those tools for licensing reasons, project picked BSD tools for their smaller code size and reduced complexity.
> But then again, it is presented here as a simplified desktop, while GNOME-based and KDE-based systems are on the images it provides.
Yeah, when I read the title here, I thought it would be about a simplified desktop environment. But they seem to just be using the standard stuff. So, I don’t think this is what they meant to say.
Simple and reliable are engineering friends.
Simple does not have to mean basic.
Take Systemd for example. Chimera Linux was to implement the same functionality and be full featured. It wants to do this with a simpler, more modular, more understandable, and more maintainable design.
>They should work reliably and predictably like every good procduct.
So you're saying they should be simple?
>unless you are a very simple person who always only does very simple things and lead a uniquely simple life
Considering the massive popularity of MacOS and iOS which all mandate the same reliable and predictable user experience on every Apple device, most people are "very simple" persons who "do very simple things and lead a uniquely simple life".
The power user paradigm of old that expose all the knobs and dials and levers there are to pleasure us simply does not appeal to the commons.
Neither iOS nor macOS especially are even preferred by the majority users so this is a bizarre argument to make.
If anything, their biggest criticism is that they're not capable enough for many productivity workflows.
For iOS there’s absolutely an argument that it’s not capable enough, but as a regular user of all three major desktop OSes I find such complaints about macOS overblown. While there’s small kernels of truth here and there much of it comes down to macOS being built around a different set of conventions than the desktop environment they’re most accustomed to (usually Windows) than inherent incapability. This is further evidenced by how it’s common for longtime Mac users have similar complaints about the Windows desktop being inadequate/incapable.
Desktop environments and user workflows are insanely personal things, not unlike clothing, diet, and music preferences but for some reason many in the tech sphere refuse to acknowledge this and try to position their preferred environment as objectively more correct/superior/etc. It’s really tiring.
> The power user paradigm of old that expose all the knobs and dials and levers there are to pleasure us simply does not appeal to the commons.
I feel that I wasted years of my life changing UI themes and colors since Windows 3 and MacOS 7, and, frankly, I have never felt tempted to do anything like that since Gnome Desktop. On the more vanilla Gnomes I don’t even change the wallpaper.
Haven't you heard that the terminal is the most simple and efficient interface? Obviously everyone should just use terminals for their tasks /s
It is conceptually very simple and it pays to learn it.
Except for most ordinary use cases. I guarantee you that, if someone downloads a compressed file, they'll find it infinitely easier to right click and select the "Extract here" option than whipping out some tar command. The linux-brain has people convinced that the layman would prefer the latter if only they would listen.
That's some windows-land programming right there. The idea that a file is something clickable, some icon with a title somewhere on the screen.
And typing 5 letters is clearly not infinitely harder than right-click and rummaging through 20 item menu to find the right action. The difference is actually infinitesimally small.
I'll hold your hand while I offer this groundbreaking revelation: the vast, vast majority of people are in Windows-land. We're even having issues with the younger generations who do everything on their phone and this don't really have any concept of file systems. Get out of your echo chamber. That means getting off of HN and start looking at how laymen use computers.
[dead]
never heard of it, but interesting setup. FreeBSD user land with Linux and Gnome/Wayland.
https://en.wikipedia.org/wiki/Chimera_Linux
Another interesting, currently ongoing, mixture of Linux world and FreeBSD is NixBSD. A NixOS fork with a FreeBSD kernel.
https://github.com/nixos-bsd/nixbsd
I came across it earlier today.
It’s not mentioned in the readme, but from a quick peruse it looks like that project also has some support for OpenBSD which is very neat.
One of the side benefits of using the FreeBSD userland is all the flags will be consistent regardless of the of program. ‘-l’ or ‘-a’ are going to do what you expect irrespective of the command. It’s nice, and I used free for along while. I still know BSD syntax better than the SystemV equivalent.
I have also used the GNU tools but the BSD ones are winning me over. Another benefit is consistency with macOS I suppose. It also uses the BSD userland.
Why single out GNOME? There are also KDE images, and of course, plenty of other DEs and WMs in the repos which you can always install.
The project founder uses GNOME and it says that GNOME is the desktop in various places including the project website and Wikipedia.
Other options are available though and there is even a KDE live bootable image. I use Chimera Linux with KDE.
For a while there was also a distro project working on the opposite direction. Debian GNU/kFreeBSD.
An official Debian GNU distribution using the kernel of FreeBSD instead of the Linux kernel.
https://wiki.debian.org/Debian_GNU/kFreeBSD
Development of Debian GNU/kFreeBSD officially terminated July 2023 due to lack of interest and volunteers.
https://lists.debian.org/debian-devel/2023/07/msg00176.html
There used to be a Debian with a GNU userland on top of a FreeBSD kernel. Debian GNU/KFreeBSD, I think.
So simplified I couldn't figure out how to even login, let alone install it.
I think if they're aiming for simplicity, start with the installer. And no, "read the instructions" isn't making it simpler.
Simplicity in implementation is not necessarily the same as simplicity in usability. Both are desirable, but if there's a trade-off between the two, then I reckon it's quite justifiable to prioritise the former.
Try installing OpenBSD then. Even though the installer runs in text mode, it holds your hand thru a standard setup, all the way to a graphical login. The basic X11 desktop is a part of the base system (as opposed to packages/ports).
The installer also supports a bunch of advanced features, like scripting an automated install. OpenBSD is also some of the cleanest and simplest code I've ever read.
Maybe it seems a stupid question, but it's an honest one: why openbsd rather than freebsd?
I tried freebsd when Debian integrated systemd. Being neither savvy or intelligent, I had trouble with the ports system and installations were taking hours for small applications. I gave up.
I've always had a nagging desire to try again, but poor health has discouraged me from those kinds of learning curves. I recently left Debian again and use Void, with Opensuse on a separate drive for purposes of maintaining some competency and familiarity with systemd (just in case...).
Anyway, I'd appreciate your input on the various BSDs. As I understand, recent funding suggests it may be growing and will remain viable for a long time, which was, however relevant or not, a concern for me.
Edit: I might also mention that I gave serious consideration to Chimera before settling on Void, for similar reasons of learning curves. Of all distros, I had determined it would either be Void or Chimera and nothing else - or BSD.
You don't have to use ports, FreeBSD has supported pre-compiled binary packages since 2014.
https://klarasystems.com/articles/a-quick-look-at-the-histor...
Thanks for the info. Time flies.
> I recently left Debian again and use Void, with Opensuse on a separate drive [...], I gave serious consideration to Chimera before settling on Void [...]
Have you tried to write down your goals/expectations? It seems like you're throwing distros at the wall to just see what sticks. I don't care for apt vs xbps or systemd vs openrc. My own goals are simple:
- I want a "UNIX workstation", overall decent for software development, need to run complex proprietary software (Logic, Compressor, Photoshop...), interface with niche hardware, develop for iOS, and occasionally play games: macOS
- I want to play a couple of games that won't run on Mac/Linux/BSD: Windows
- I want a simple system that I can understand and actually enjoy using: OpenBSD, Alpine
> purposes of maintaining some competency and familiarity with systemd
You can also just use a VM or a homelab box. No need to care for your personal device, as long as it fulfils your stated goals.
> why openbsd rather than freebsd?
I think OpenBSD is to FreeBSD what FreeBSD is to Linux, and then some more. I'd say these are its most distinguishing features (compared to almost every other OS):
- User-friendly - assuming you have some experience with the command line, can follow prompts, read documentation, etc.
- Excellent documentation, check out the FAQ: <https://www.openbsd.org/faq/>; manpages: <https://man.openbsd.org>; etc. Even compared even to the FreeBSD handbook, which (while a step up from most Linux distros) isn't quite on the same level.
- Secure by default: it is the upstream project for OpenSSH, LibreSSL, PF (notably forked into macOS and FreeBSD), and some others; also: privilege separation, W^X memory, (K)ASLR, doas, signify, pledge, unveil, kernel/libc random relinking, syscall origin verification, etc. Best part, you get all of that for free and you don't even notice. There's no "hardening guide".
- Truly self-hosting - the base system has everything you'd require to comfortably develop the base system, including X11, clang, mg (an Emacs clone - if you're not a vi fan), tmux, etc. You will probably install packages anyway, to get your preferred tools, but the base system will still be there.
- Innovative, not just technologically but culturally: cvsweb was the first of its kind - the source code, including all commits, could be anonymously browsed via the web, in (IIRC) 1996; most contemporary projects would just occasionally throw tarballs over the fence.
- Simplicity. It is not just the manpages or the source code; OpenBSD meticulously rips out or refuses to accept complex systems (sudo, gpg, seccomp) and instead introduces simpler interfaces (doas, signify, pledge)
I wish I could live in a world where this is all I need, but unfortunately, OpenBSD brands itself as a "research OS", and doesn't provide any accommodations such as binary compatibility between releases, Linux syscall emulation (which FreeBSD does), and so on. Even vmd(8) is a recent addition.
Hell of a reply. I'd definitely buy you a coffee for it.
Thanks for all the considerations.
PS: regarding throwing distros at the wall; I used the same installation of Debian for 10 years. It was my workstation and everything else. But its direction isn't mine anymore, and I gave quite a bit of thought to my choice, ie Void. FreeBSD would be in addition, not a replacement. Also, my primary needs are far more primitive than yours or what you mentioned. I mostly need something I can trust, that doesn't change for the sake of change. That and for it to be able to do basics, eg python without dependency hell, video editing, office, music/editing, caveman AI, and such. Hardly more. Oh, and freedom. That's really important.
I've also been on Debian for 15 years, before switching to macOS. I had a dozen very good reasons (incl. I've finally been doing the work I've always wanted to), but the "change for the sake of change" seems kinda Apple's motto nowadays, and yes it's hella irritating. Still - for me, macOS beats any other OS in that it fits most of my needs almost exactly.
And when I want a Mac that's much simpler, I also have a TiBook! <https://www.rollc.at/posts/2024-07-02-tibook/>
> Oh, and freedom. That's really important.
The only true freedom is to do what you desire. For me, it's e.g. playing StarCraft. Yeah I can make it run on Linux under Wine. No, figuring out why it's broken again is not my idea of free time ;)
You might like Chimera Linux. It also “refuses to accept complex systems”, except it is still Linux and so does not have the limitations of OpenBSD.
Chimera Linux also uses doas. The Turnstile project is a good example of Chimera Linux rolling its own when no suitable alternative is available (to replace Systemd in this case).
I see your points, but Alpine already checks most of these boxes, so for me the cost of switching outweighs any expected benefits. (I don't believe in "just trying it out" in a VM, because the real issues crop up only when daily driving.)
I assumed this was a post about ChimeraOS, a gaming focused Linux distro, but it seems this is talking about a totally different Linux distribution with the same name.
Chimera OS: https://chimeraos.org/
Chimera Linux: https://chimera-linux.org/
From the FAQ (https://chimera-linux.org/docs/faq#what-about-chimeraos)
“The system also has no relation to ChimeraOS, besides the unfortunate name similarity. ChimeraOS used to be called GamerOS and renamed itself to ChimeraOS later; however, at this point Chimera Linux was already in public development with its name in place.”
It's the userland that always sends me back to Linux. I've tried BSD a few times, but there are enough gnu flags in my muscle memory for command line utilities that it's painful.
This is an honest question. Do you have any examples?
I have always been a GNU user but have been using Chimera Linux. I ran into differences with ‘sed’ and read about a regex difference with “find’. Really curious what flags other people are using.
one sticking point for GNU users can be that BSD utils are more strict about positional order of e.g filename parameter and input options. rm ~/foo -rf || rm -rf ~/foo
I was forced to get used to BSD syntax when I switched to MacOS but now I prefer it...
Thank you for the answer!
Even as a GNU user, I never would never have thought to put filenames anywhere but the end. I mean, with wildcards, it could be any number of filenames.
You are right though, that is an important difference.
It’s unnatural to put the flags after the command. But sometimes I remember I’ve to use a flag after I’ve entered the input; other times, I notice a flag’s omission after executing it, and Control + P puts my cursor at the end, after the input.
A pleasant surprise - the KDE version works "out of the box" on my 2015 MacBook 12" (A1534). This is an unusual machine, and Linux distros tend not to support its hardware out of the box. Chimera even sees the internal SSD, traditionally a stumbling point to installing Linux on this machine.
It has been amazing to read in this thread how many people seem to think that “nothing new in the UI” equals “nothing interesting about this distro”.
It really highlights how the relationship with computers is changing.
The goals of Chimera Linux have little at all to do with the UI. In my view, Chimera Linux is highly innovative and disruptive. Yet, some people can look at it and genuinely wonder what it does differently or what the point of it is at all. Fascinating.
> Chimera supports many different configurations, leaving the user free to carve up their disk as they please — but does not support having /usr on a separate partition.
On systems with both a hard drive and ssd (since on most distros /(s)bin and /lib* are sym links into /usr, most binaries are ultimately in /usr), my recent preference is to have the hard drive as / and ssd as /usr (then sym link or bind mount there as needed for things that could use the speed, like Steam). Am I the only one who thinks this way or am I way off?
I don’t think of ssd and hdd for different parts of the filesystem. Rather / is a zfs/btrfs/bcachefs pool, and the ssd is added as a read cache to the pool
Edit: it can also be a write cache but that’s more tricky, usually with a battery backed hardware raid it’s fine
> Chimera's recommended desktop is GNOME
But that's the opposite of a simplified desktop
GNOME? It is super simple. I have used Fedora/GNOME on my elderly family members laptops for many years now, and they just get it. Even the ones who came from Windows 10. The hours I spend on support has dropped significantly. Windows is such a hassel, and its desktop design philosophy is just not that great for casual users.
IMHO, GNOME is not simple, it's dumbed down. They've obviously tried to copy Apple where it suited them, but missed on a whole bunch of important details, like a unified menu system, a powerful terminal emulator, desktop icons (omg), while badly aping all of the worst parts - from requiring 4 clicks to shutdown/reboot (where macOS requires 2), thru a mostly useless top bar that steals the real estate from browser tabs (Fitt's law), to asking SDL to link against libadwaita to draw window decorations. And this is the worst part, they not only do not want to accommodate their users, but also ignore the developers - those who wish to integrate with and therefore empower the free desktop ecosystem.
Apple can get away with all of that because they're a trillion dollar company, but unlike Apple, the power of the open source community doesn't stem from an unimaginable pile of cash, but from interoperability and cooperation.
I tried to use GNOME for about 8 months but there was just too many WTF moments (one caused by Ubuntu's own dock extension). There's plenty to love about GNOME but missing features, bugs, and design/usability issues makes it feel like beta quality software.
A few months ago, someone wrote a blog post[1] cataloguing many of these issues. One thing not mentioned was the lack of a caps/num lock on-screen indicator, this is a feature that is present in GNOME 2, MATE, XFCE, Cinnamon, KDE, and Windows 7/8/10/11 out of the box and toggleable in their settings window. The only way to gain this functionality in GNOME is through a third-party extension. Many laptops continue to be manufactured without a caps/num lock indicator on their keyboard, it's insane this isn't a supported feature in GNOME.
I don't get the feeling that GNOME team has ever implemented accessibility research in their design choices. For those with disabilities, GNOME is unnecessarily difficult to use [2][3].
[1]: https://woltman.com/gnome-bad/
[2]: https://old.reddit.com/r/linux/comments/s3vvot/state_of_acce...
[3]: https://lobste.rs/s/3bdlpn/thread_on_deterioration_accessibi...
I can't think of a single thing in Gnome that is somehow better than Windows, in terms of usability for beginners.
My elderly family members beg to differ ;)
So far, my experience with people above 60 is that they quickly understand how to find and navigate between the apps they use every day. The ones coming from Windows adapt suprisingly fast, and are very pleased that the fullscreen popups and forced updates are gone. Shortly before I moved them over to Fedora, they started developing a fear of using the computer because they never knew when it would 'lock them out' because the screen was filled with a fullscreen Office365 ad that had no obvious exit button.
It's also not really anything original and new? Not sure what the overall plan is here
I like that they support RISC-V, thus showing an understanding of the direction of the industry.
Why not just use a BSD?
The obvious answer is to build on the Linux kernel. That means more extensive desktop hardware support. That means better desktop software compatibility. That means robust support for OCI containers.
But Chimera goes beyond the BSD model in several ways. For example, it aims to bring the Systemd feature-set while avoiding Systemd.
It also uses pipewire and Wayland.
As mentioned elsewhere, the Chimera Linux founder also found the FreeBSD packaging system to be lacking.
Chimera Linux also aims for stateless /etc and /var.
There is a lot more to Chimera Linux than the userland.
Ok, so why use any BSD components then?
Is this a real question? Instead of justifying the choice to use something good without taking the bad bits too, let me pretend you asked “what are the advantages of using the BSD userland”.
I am not a Chimera dev and they have answered this but here is what I understand….
You could argue that the GNU utils are bloated and over engineered. But that is just an opinion. As the Chimera devs point out, there is a “chicken and egg” dependency problem with the GNU utils when trying to build from source and / or in a container. The BSD utils solve this problem.
At the same time, the BSD utils are more complete and powerful than, for example, Busybox or other alternatives to the problem above.
The BSD userland solves one of the problems that Chimera Linux is trying to solve. The BSDs do not solve many of the other problems and so using everything from BSD is not an option.
I will point out that Chimera Linux and FreeBSD both use Clang/LLVM as the system compiler. Chimera uses it for LTO and certain security features.
Linux driver support. There’s a ton of wifi drivers and the like that are supported under linux but not under BSD.
Obscure driver support doesn’t seem like it fits with a “simplified” Linux but maybe…
The list of hardware supported by the Linux kernel is not exactly “obscure”.
Wifi is not “obscure”.
considering it's not a bsd nor it's trying to be like one, that question doesn't make a lot of sense?
Definitely going to take this for a spin. BSD has been appealing to me for a long time, but the driver situation is wanting. In theory, a more integrated BSD like experience but with Linux kernel/drivers? Interesting!
I appreciate what the Chimera authors are trying to achieve, but I would never consider going back to a distro that doesn't support atomic upgrades and seamless rollbacks, ala NixOS, Guix, etc. Packaging issues and incompatibilities are inevitable and impossible to prevent. Giving users the peace of mind that their system can always be reverted to a known working state is priceless, and something all operating systems should have.
Making the system configuration declarative, reproducible builds, etc., would also be a bonus, but I wouldn't consider those hard requirements.
I realize snapshots are a feature of some filesystems, which partially addresses this, but I would rather have this feature at the OS level.
Fair enough. It is worth talking about how the Chimera Linux package manager works though.
There is a file, /etc/apk/world that lists all the packages you have explicitly installed. When you add and remove packages, all it really does is change this list. Then it runs the solver and installs the package versions and dependencies required. You could move /etc/apk/world to another system and it would result in exactly the same set of packages between the two.
Replacing /etc/apk/world to any previous state will “roll back” the system to that point.
The package manger is transactional and apk commands either entirely succeed or fail. When I was testing Chimera in a VM, I ran out of drive space during an install of dozens of packages. The system was left in the same state as it was in before I ran apk.
Cports itself is very declarative. Also, unlike most distros (including Arch) there is no separate install and update step. Other than the kernel, and maybe a couple others, nothing runs after the package install.
Finally, Chimera Linux is aiming for stateless /var and /etc.
Chimera Linux may be closer to what you want than you expect.
That's good to know, thanks.
> Replacing /etc/apk/world to any previous state will “roll back” the system to that point.
That's not quite what rollback means, though.
Say that a new version of a display driver crashes my display server. With NixOS I could reboot into a previous configuration and have a working system in no time. If I have to keep track of /etc/apk/world changes myself, and boot into recovery mode or chroot to fix this manually, it takes time, effort and frustration I'd rather not have to deal with.
Stateless /var is a good idea, but I'm not sure it would address this problem.
In any case, I'll definitely keep an eye on Chimera. We need more distros that are not derivatives, and try to do something different. Wishing the team the best of luck, and hoping the project succeeds!
Had a go at booting this in a VM. The ISOs apparently have some issues, meaning you can't just boot in a VM and have a poke around...
It's stance on keeping it's own reimplementations of parts of systemd as a dependency sadly rules it out for me. It would be a great choice for a linux jail on freebsd.
I hope Google will make Android more desktop friendly so it becomes the real alternative to the current status quo.
I don't. Google's Android support has just slowly eroded away all the Linux benefits with layers and layers of user hostile "security" without the benefits (i.e. backing up my phone to Google doesn't mean a new phone will one-click recover itself to how it was).
We could have something like android in the linux world if manufacturers did the minimum to support Linux (which already accept binary blobs). But they also want to push their (bloated) apps alongside the driver.
They have ChromeOS Flex which runs on normal PCs and works fine.
I hope not
Simplified desktop doesn’t very clearly map to “very text oriented, barebones Linux UI“.
I was expecting something more like the ststem76 Pop OS kinda thing when I read this.
Niri would be a simplified desktop to me
https://github.com/YaLTeR/niri
I really wanted to switch to this but unfortunately it doesn't have support for Xwayland scaling :( in case youre using 4k, On Wayland only KDE and Gnome seem to have gotten this working.
There is also PaperWM https://github.com/paperwm/PaperWM it works pretty well. I would be using it if i didnt switch to Cosmic.
If you're on KDE check out Karousel. It's a KDE Kwin script that does this type of scrollable tiling right inside KDE. Works great on my ultrawide monitor.
https://store.kde.org/p/2045724
Niri is my daily driver. I absolutely love it. Couldn’t switch back to any other approach to window management.
Wow... love it, thanks for posting.
Cheers, that looks nice.
I think it would be better to say “simplified desktop Linux”. The project is more about the Linux “plumbing” than about opinions about the desktop environment.
Chimera is not even out of beta and it already includes GNOME, KDE, XFCE, LXQT, and others. The founder uses GNOME but Chimera users do not have to.
[flagged]
[flagged]
> Surely, 2025 will be the year of the Linux desktop!
I don't know if that joke was more funny in the late nineties when Linux was basically nowhere (I was already using Desktop Linux back then though) or if it's more funny today, when Linux powers the entire real world.
I mean: Linux powers 500 of the Top 500 supercomputers, billions (if not tens of billions) of devices, rockets (SpaceX uses Linux extensively), servers all around the world, routers in every single household or near that etc.
Basically: Linux powers everything besides the desktop.
Somehow I think there's some irony in thinking that joke is still funny in the same way it was funny in 1998 or so.
It is a reoccurring joke, aka a meme (before internet memes were internet memes). If you liked it, it builds on repetition and nostalgia. If you disliked it, you won't like when repeated. On top of that, whether you liked it or not it may get boring.
Nowadays, it is out of date. Reality surpassed it: 99,9%+ of all smartphones is powered by either Linux or Unix (macOS). Are these desktops, with a pointer device? No. But they are GUI clients with touch UI. The Windows version? Too little, too late.
Same counts for Linux desktop: too little, too late. If it ever wants to overtake Windows or macOS it has to be substantially better. This is why some FOSS people dislike efforts like Cygwin, SFU, and successors (such as WSL). It allows Windows to use unique utilities which were tailored for FOSS OSes.
It's the "...on the desktop", i.e. that the general public would have a stated preference for Linux because it served them better than proprietary OSes, putting their makers out of business, that makes it funny. Linus even said "Really, I'm not out to destroy Microsoft. That will just be a completely unintentional side effect." in a NYT interview back in '03. Much like "640k ought to be enough for anyone." and "There will be a market for only 5 computers in the world." (only in this case Linus is actually on record saying it), it's a testament to the overweening hubris of the early FOSS movement.
The things that you list matter only to a small number of nerds and the general public wouldn't care a bit if they were all replaced by FreeBSD, Fuchsia, or TempleOS tomorrow. Even for the consumer applications, Linux is being primarily chosen just because the hardware manufacturer didn't want to pay money to license another OS to benefit their profit margins as much as anything else.
So, yeah, the joke is still funny and it looks like it will be funny for the next few decades too.
If it was "born from unhappiness with the status quo" why not join forces with Void Linux?
Void Linux is part of the status quo they were unhappy with.
Chimera Linux started off as a rewrite of the Void source packaging system. One of the complaints was using shell scripts for package templates. Chimera cports uses Python. Cports builds the binary packages (for apk 3).
TBH depending on Python also isn't the best choice - it has become a huge and complex language, goes as far as optionally embedding another full language (Tcl, for Tkinter), all the async stuff, kitchen sink standard library, etc. I don't even know how much effort it would require to bootstrap.
I would personally go with something like Lua - small, trivial to embed, multiple implementations to pick from, etc. It would fit exactly where the shell falls short.
it's absolutely the kitchen sink standard library and "complex" language that makes it worth using, because 1) it means no additional dependencies and 2) the language is expressive enough to let the template syntax remain simple and well-abstracted
have you ever tried bootstrapping python? because it's really not a big deal (the only hard dependencies are a C compiler, libffi, and zlib, though distros also add libedit or readline, expat, sqlite, bzip2, xz, and openssl, all trivial dependencies present in more or less any system); tkinter is pretty much always compiled and packaged separately
meanwhile with lua you'd need many additional modules, and the templates would be verbose and far less readable
a non-exhaustive list of stuff python provides that would need to be provided separately:
1) subprocess management 2) path parsing (and normalization, manipulation, etc.) 3) various filesystem operations 4) shell command lexing 5) globs and fnmatch patterns 6) json 7) date/time handling 8) filesystem advisory locks 9) temporary file handling 10) http client (for sources fetching without external subprocess) 11) termios and ptys (for isolation of controlling terminal) 12) regex 13) hashing 14) topological graph sorting 15) cmdline argument parsing 16) readline functionality 17) thread/task pools
and so on and so on
Don't get me wrong, I love Python, and I've been a user since the 2.4 days. You brought up plenty of good reasons to use it - and my own rule of thumb is to rewrite a shell script in Python once it goes past 100 lines.
My point is that Lua-the-language provides similar expressive power to Python-the-language (including basic stuff like lookup tables), while remaining smaller and simpler to understand than the POSIX shell. It's also simpler to vendor than Python, and you probably want to vendor this kind of stuff to avoid the pain of bootstrapping / circular dependencies.
Now I haven't built a build/packaging system like xbps or apk (few people did), but I've been in charge of devops/releng at work for a decade - and if I learned anything at all, it's that it's better to start with the simplest tool that can do the job, and slowly add what you need. E.g. I wrote a tool in 2016 to replace Ansible in my team, and all of this time it's just been doing its job: <https://github.com/rollcat/judo>. Turned out nobody actually needed the kitchen sink.
note: i've been doing lua for two decades and am well familiar with the language itself as well as its inner workings (far less so with python), and i've been building distro tooling for almost as long, and no, definitely not
I do enjoy Lua, what I miss the most is a more complete and direct interface to POSIX - possibly as a compile-time option, to accommodate the existing embedded use cases.
What would you say is its biggest shortcoming?
Have you ever tried upgrading python on a machine with a large codebase that you need to maintain?
sure, in fact i updated an entire distro to a new python major version several times
cbuild started iirc with python 3.8, currently it requires at least 3.12 due to some features (it will stay on that minimum for a while, though newer versions are always supported), updating it has always been seamless though
The creator used to be a Void contributor.
The were the PPC maintainer. Which is why PPC is a tier one arch for Chimera Linux from the start.
Interestingly enough there are almost zero reasonable PPC computers one can get.
At the PPC prices, you got to have a really good reason to want a PPC desktop. The TalosII cost nearly $10k for the base 2X 4-core Power9 CPU version, and another $5k to upgrade to the 22-core CPU version.
https://www.raptorcs.com/content/TL2WK2/intro.html
I see, thanks. https://chimera-linux.org/docs/faq#what-is-the-distros-relat...
And this is why, eventually one gets tired of all reboots, and endless variants on Distrowatch, get a computer at BestBuy, Mediamarkt, Dixons, or whatever is the local chain, run GNU/Linux on a VM, and be done with it.
The Chimera Linux team strives to downplay the lack of GNU because they do not want the politics.
For me, one of my favourite things about Chimera Linux is that it is both everything I want from Linux and 100% immune to being called GNU/Linux.
Chimera Linux: no Glibc, no GNU utils, no GCC, and no Systemd. As a user, I am finally free.
We should consider “freedom from being labelled GNU/Linux” as “the 5th freedom”.
Basically, had it not been for AT&T lawsuit, or Microsoft losing interest with either Xenix or Windows NT/POSIX subsystem, and Linux would hardly matter.
You're pointing out, in other words, that UNIX/POSIX is what matters, regardless of the package.