I'm surprised that a Wenux fan boi would only mention the ChromOS termina VM with the Debian LZD container and leave out Chromebrew.
Linux is now a little more than three percent of global desktop OS market, excluding the just-over four percent that is ChromeOS. Which is also Linux, but the wrong kind of Linux. Web server statistics aggregator Statcounter announced last week that as of June 2023, Linux accounts for three percent of worldwide desktop …
...why have I never been able to get it running in a Qemu/KVM virtual machine? To install ChromeOS Flex, you had to tell it what the hardware was, and only some specific machines would be accepted.
Or has that changed?
If it's open source, I suppose I could build a version from scratch.
xorg.conf and the monitor timings was the only example I've personally come across in a 40 year career in technology were a configuration might destroy hardware. Fortunately, I never did but there were a few occasions when I didn't like the noise coming out of a monitor and quickly unplugged it.
Not Linux, but startx brings back some memories.
Mid 90s, during the first year of my degree. Wednesday afternoons, in the big computer lab. Rushing to the lab in hopes of getting onto one of the few Sun Workstations we had because we had a whole afternoon of being taught Unix based stuff, and the Suns were a whole lot more pleasant to use than the PCs next to them, which could connect to the Uni's Unix systems via a package called "Exceed".
Exceed, which still seems to be around, was a Windows based X client. Exceed wasn't bad by any means, it's just that the Suns had better keyboards (although with a crap mouse) and they had proper Sony Trinitron monitors, rather than the generic no name PC monitors the Uni user. They were also 17 inch, as opposed to 15..
>Mid 90s, during the first year of my degree. Wednesday afternoons, in the big computer lab.
Late 80s but we ran to get one of the few amber screened serial terminals with the nice separate keyboards rather than the ancient green screen VT100s
25 students running of a Motorola 68020 powered Sun 3 with less compute than a Pi-zero
Technically, Exceed was a X server. xeyes and all the wonderful X applications were X clients. This always seemed to confuse people. I think this was because people had this idea that the system that was "close" to them was where clients ran and the far away system was the server. If instead, you thought about whether a piece of software was initiating or accepting a connection; it might have been easier for people to remember. So you might use a telnet client on your local PC to connect to a telnet server on a remote system, set an appropriate DISPLAY variable value in the remote shell session, and then run an X client on that remote system to connect back to the X server running on your PC. So software running on the same machine might be either a client or server depending on exactly how it was being used.
I don't think you can still get Slackware on floppies, as that would require about 2,500 of them. But it is a great way to learn Linux. There are good tutorials on installing it, but make no mistake - you will be formatting partitions from a command prompt. You will learn a lot about Linux if you install it.
My first foray into Linux was on a 3-day weekend, using a 1-floppy slackware router-bootstrap config on a 486 pc with a linux-blessed 3com ethernet card connected to a 256k ISDN line. It was a steady string of downloading packages/libraries, compiling, reading man pages, and repeating. 30 hours later I had consumed 2 pizzas, 6L of soda and had working GUI with audio.
Very educational process and I think more people should do it just to understand the hardware stack.
“ The Linux World TM is defined by The Front of Linux Power Users!”
.. or so say the Peoples Front of Linux.
… and also the Popular front of Linux.
Any you wonder why people smirk when you try to upsell Linux v’s Window’s, MacOS etc with this petty Football Team or Religious. Zealotry.
I think if you look you'll discover that the git in question was the bloke who revoked free use of Bitkeeper, forcing Linus to build a new distributed revision control system that could handle the needs of the fine folks maintaining the kernel. From scratch. In a hurry. git was self-hosting in 4 days, and went "live" handling the kernel in about two months.
It wasn't until much later that Linus claimed that he was the git in question ("I name all my projects after myself."), but the folks who were in the trenches at the time know better.
" ... "The Linux world"? What even is that? ... "
Anything that uses the Linux Kernel is part of the "Linux world".
" ... Who decides? ... "
YOU do. You get to decide. Quite a concept isn't it?
" ... I thought the whole point of Free Software was that you get to use it however you like; even ways other people don't like. ... "
See, you knew the answer to your own question. You get the point.
You have any actual *proof* that, say, the kernel is spying on you? How is it doing that? Maybe they modified bash to report back to homebase?
Given that normal ChromeOS users spend their entire time within the browser and extensions, they don't need to fiddle with anything else, do they?
And note that I did not suggest another Chromium based browser, so no claiming they are all using the same core engine and spying that way.
"... slurp, slurp and slurp..." I agree, but to be fair, just like Windows, Apple etc. Google do put their significant improvements back, they have to.
My experience: I have managed computer systems for over 40 years, Chromebooks have been my go-to device for 10 years now, I have used many dozens in the workplace too, they are not throwaway, they have negligible administrative maintenance. I still love them.
I have a super duper 16" MBP but I also have a Pixelbook, which I bought in 2017 and I still use it most days. It's secure. It's light and it's compatible. If all I'm doing is surfing the web, it's the perfect device. If I want to play on my fpv drone sim, I need to use a powerful machine. That is not a big proportion of my time and my Chromebook still does what it should very well after 5 years. That's well over twice as long as I keep my phones for.
Glad it "works for you", some of us would like other browser choices.
(and yes, if you install all the other components you can install one, but with a fraction of the effort you could have booted a portable live image with a lightweight actual OS and done the same thing.)
That said, once you accept the yoke of Google Chrome, it really doesn't matter what you run it on.
But the cheap Chromebooks really are trash, every few years some clown pushes them as an alternative to real machines, and wastes a few days our lives shooting down the idea that a 200$ warmed over netbook can replace something like a surface or a macbook. Eventually I give up reasoning with them, and feed them to the business office and tell them they want to take away their adobe creative cloud.
Which is kind of unfair because I actually WANT to, but they aren't bright enough to realize you can't run the actual software they want on a Chromebook.
"Chromebooks are the Netbooks of the 2020's. Designed to be used a bit and thrown away"
I have a netbook from the Windows 7 era I've not thrown it away because I wanted it to use, and still do use it, when the need is for something physically compact. It still works because, of course, I'm running Linux on it.
Plastic case designed to be dropped, keyboard that is junior school drool proof, no fscking touchscreen
No I don't use it for doing protein folding while international jet-setting from my business class seat - but for watching youtube or browsing reddit on the settee it's great
Hmmm, that's funny... my "travel computer" Chromebook seems to have some aluminum in it and definitely appears to have a 1080p 180-degree folding touchscreen so I can watch cat videos on airplanes! I'm not a huge fan of the keyboard, which is why I have a folding Bluetooth thingy that travels with my Logi trackball mouse, but the system can (and does) drive a second screen if there's one handy and definitely does what I need: the usual travel stuff via websites, and establishing a VPN and Remote Desktop Connection so all my data remains Not With Me.
I'd prefer it if someone didn't steal it on my travels right now, but in a year or so, I'll replace it with the an equivalent at that time. In the meantime, if I do lose it or it gets lost, its cheap enough that I'll just grab another from our IT department (called "Bob") and carry on...
Could it be better? Sure... the keyboard is gray and backlit so you currently can't read the white legends on the keys, which is pretty daft, and it's a 14in machine so the keyboard is smaller than I'd like (next time I'd agonize more over the 14in size/15.6in keyboard/numpad tradeoff!). But it does the job as a remote terminal, and I don't need (or want to pay for) more than that.
"all my data remains Not With Me"
If you're a traveller this is a good idea but if I were a traveller using it for anything sensitive I think I'd want to have a choice of where the data remains with the device completely amnesiac about where it expects for find it. That way any inquisitive border official can be presented with an innocuous target on Google whilst the confidential stuff remains on a private NextCloud server I'm not sure whether the Chromebook does that.
"It still works because, of course, I'm running Linux on it."
Same here. Some crappy old Asus EeePC I was gifted. Seems quite happy with Linux Mint[*] and firefox with 20+ tabs open at once. It's a bit iffy with YouTube videos above 480p, but maybe I need to spend some time tweaking it or maybe it simply can't cope well with that. But for general/casual browsing while watching TV, etc, it's mint!! :-)
[*] Mint was the first distro I tried installing that found all the relevant hardware, including the camera, direct from install, so I stopped experimenting at that point. All it had to do was work acceptably so the 2 or 3 previously tried distros simply got overwritten until one worked without any post-install messing about.
I also have a Win7 era netbook. It's on its third HD, and running Linux as an online file server 24/7. It wouldn't do as a general purpose computer, and the occasional web browse is sort of slow, but otherwise it doesn't require much attention, it just works. I suppose a RPi would use less power, but power consumption compares favorably to all the other computers I've got running.
My opinion is quite different. I think Chrome OS certainly does count as a variant of Linux. I do not extend that to Android, because there is no compatibility between it and any other Linux, but since you can run Linux programs on Chrome OS and it meets the other requirements, I think it counts. If I were interested in having a rating board of how many Linux installs on non-server devices exist, I would count Chrome OS.
The reason I don't support Chrome OS is that the number of kernel and userland installations is not of much interest to me. What is of interest is the freedoms that non-Chrome-OS Linux distros tend to provide, which Chrome OS didn't and Chrome OS Flex only sort of does. I dislike Chromebooks for various reasons, including the designed-in dropping of security updates for no good technical reason, something very contrary to the ethos of most other Linux-related projects, especially the kernel itself. I don't like the locked bootloaders that the hardware often bring. I'm not very concerned about the year of the Linux desktop, as it's not going to happen and I don't really need it to, but to the extent that I would like it, I would like it for what it means for users, which means that Chrome OS wouldn't count.
I'd agree that ChromeOS is a Linux variant: I went looking for an IMAP client that I didn't detest, couldn't find it in the app store that mysteriously does exist on my Chromebook (despite what the article implied), so I enabled the Linux mode and installed Geary which behaves exactly like Geary so I have no reason to believe a Chromebook in Linux mode is not Linux in any meaningful way!
"What is of interest is the freedoms that non-Chrome-OS Linux distros tend to provide"
And this is actually the main thing holding Linux back from being a great desktop experience for most users, and why Chromebooks have succeeded where $otherLinux hasn't.
Normal, every day, people do not want an enormous array of choice. They generally want a curated experience. Look at the phone ecosystem - iPhone and Android are both very rigid in how they work. Sure, you can jailbreak Android and fiddle more, or do some sideloading etc, but the normal experience for an end user is simple.
Having a choice of a thousand distros, and then the sub choices within them (desktop environment, whether to use something like snap or flatpak, etc...) confuses people.
Freedom is great, sure, but a lot of people want simple, not freedom. (And that appears to apply to all aspects of life it seems).
"Having a choice of a thousand distros"
With an Android phone you have a choice of a thousand apps, all wanting to slurp more data than they need. What a choice.
The reality of Linux desktops, of course, is that for a daily driver you have a choice of a very few distros from which you can pick one that suits you best. You can ignore all the projects that either (a) all offer the world's smallest footprint or (b) aim to completely revolutionise the desktop in particularly odd ways.
This, of course, is still more choice than Windows which only offers two choices: take it or leave it.
I have a laptop with W10 and Devuan on it so I can compare the two. If I have a quick job to do I could fire up the Devuan, do the job and be closing down in the time that W10 takes to spin little dots and then display the login background while it has a think about displaying the actual login prompt over the top. My W10 choice is inevitably leave it.
Not really the same thing. The app store is a curated experience again.
The choice of the entire OS is a different ball-game, as people simply don't have a frame of reference for the things they need to choose from. Having multiple distros, multiple display managers, etc is something no-one has ordinarily had other than in this specific environment.
When they've bought a PC in the past, it came with an OS with a single desktop type. Or a Mac. Or a tablet. Or a phone.
And those choices then introduce complexity. Why doesn't this app look the same as the other ones, when they choose a GTK based app over a KDE based app etc...
Then throw in configuration. There's millions of configuration choices in Linux. Only a minute fraction are revealed to them via GUIs, and you end up being sent to configuration files to edit text instead. Great for flexibility. Great for freedom. Horrific for user experience.
I've been using Linux since 2000. The basic underlying user experience hasn't changed in that time. The community hasn't changed in that time, with gatekeeping rife still (just look at this article). The only thing that has changed is that some aspects of the system work out of the box, on some devices.
Mostly, they want something that works! And too often that isn't Linux. Though Windows has some huge issues now too, the base os doesn't just not give you a screen, or disable the keyboard, or insist on a degree level education before formatting the drive to use.
Yes. That's very much the case. I tried out Manjaro Linux on my desktop PC at home. I suppose I'm a little "niche" compared to an average user, as I have 2 monitors (a 4k screen and a 1080p screen), running on a Nvidia 4070 Ti. I was quite surprised to find how poorly it handled it.
Issues with scaling/DPI, changing things per guidance online broke the GUI in total forcing me back to editing text files on a terminal screen.
If this had been Windows, it would've had GUI out of the box, without needing to fiddle. Changing DPI scaling is a simple dropdown and didn't need me to know about the specifics of my display manager, or which driver I was using, or which desktop environment etc...
> I was quite surprised to find how poorly it handled it.
I have tried to cover this -- e.g.
As for Manjaro...
We find glitches like this worrying. A simpler, easier Arch Linux sounds like a good thing, but we feel that significant issues like this shouldn't appear in a point release. Manjaro is, as it says, easier to install and more beginner-friendly than Arch, but this distro still isn't one we'd recommend to a rookie.
That conclusion is me being as diplomatic as I can.
I've lost count of the number of distro's I've used over the last 20+ years. It just surprises me that a distro that styles itself as a user friendly one has such basic flaws. The version I tried was the KDE one. To a normal user trying a distro for the first time, it would be a poor introduction.
Its a shame really, as Manjaro has worked nicely for me on other systems in the past.
Oh? I started my new Dell PC a couple of years ago. It set up encryption for the drive, upgraded from Windows 10 to Windows 11, rebooted - and asked me for the encryption key.
What encryption key?
To their credit, Dell people spent two hours on the phone with me trying to break into my new computer. We went to the Microsoft website to recover the key, for example: It literally said "There's nothing for you here". (Who writes Microsoft's error messages?)
Finally they gave up and provided a phone number at Microsoft, but that person just wanted my credit card number before they could help me. I think not.
Just me? Nope. One of my students at university asked for my help. He had bought a Dell computer and it had encrypted the drive and upgraded to Windows 11, too, locking him out. I told him to return it to the local store as defective.
Just Dell? Nope. The university computer I was given came with a one-time login to allow me to set up a local boot password. However, Windows grabbed control as soon as I entered my credentials to install mandatory patches and then rebooted, giving me no opportunity to set a boot password!
Am I just unlucky? Perhaps. But I believe the real culprit is that people tolerate Windows because it's all they have used. They assume all desktops are like that, and since it runs the applications they learned early on, they stick with it. Or as many people I know like my brother who can get away with it, just switch to iPhone only.
Just my $0.02 - but I've paid far more than that in my dues! :/
"Normal, every day, people do not want an enormous array of choice. They generally want a curated experience. Look at the phone ecosystem - iPhone and Android are both very rigid in how they work."
I have to agree you on the general concept that users don't want as many choices as I do, and disagree about Android proving it. With Android, you have lots of choices of how your phone's going to work, but unless you're going to make those choices with ADB and sometimes even if you're willing to, they are choices you can only make once and then you're stuck with them. I refer to the different versions of Android made by every different manufacturer. Google uses plain Android, since it's them who are writing it, but they also have some Pixel-specific features. Other manufacturers have their own versions of things. The launchers will work differently based on what they wanted. The built-in applications will not be the same. This has gotten to the point that reviews of Android devices tend to spend a while on the manufacturer's software and some people have strong opinions for or against one manufacturer's version of Android without having the same feelings toward others (I'm mostly thinking of the relatively strong opinions of Samsung's software at various levels). Anyone who does research before picking an Android device will hear of these differences and have to decide, for example, whether they want a stock or close-to-stock version or whether some customization is acceptable.
This isn't even thinking about the kind of things that I do. The length of security updates or the speed of feature updates is also different, but I'm guessing it's not commonly explained to buyers who aren't explicitly looking for it. Similarly, I don't expect any nontechnical buyer to care about unlocked bootloaders or rooting. I'm only referring to what you see when you unlock your phone, and there are significant differences.
In a nutshell, a desktop OS only gains serious traction when it is mass-marketed as such, and comes preinstalled on your machine. That's the case with Windows, Mac and Chrome-OS. It is not the case with any other desktop Linux.
Non-techie users are never going to change the OS on the machine they bought (assuming they even know that the option exists - which the vast majority don't). And why would they, when what they get works according to expectations*?
*which may be low, but that's not the point here.
ChromeOS is designed to be safe to give hordes of rampaging 10yos, and still be supportable by the sort of sysadmin who will look at '60k/yr in a union shop (at least in the states the only place you find IT unions is local govt - and they are epically parasitic) and think 'where do I sign'....
For the US the overwhelming majority of Chromebook sales are to K-12 schools, and it is the appeal of a totally managed, super locked down laptop that sells them....
"Chromebooks are the Netbooks of the 2020's. Designed to be used a bit and thrown away which is not what we should be doing."
Hate to tell you this but this applies to pretty much all computers. They aren't designed to last exceedingly long times. Laptops are usually designed for a 3-5 year life span. Sure, some last longer, but they aren't designed to. Where would the profit be in that?
Chromebooks are designed to get 8 years of support from their original model release date (not from when you buy it).
The idea that ChromeOS is a "fake linux" is, well, nonsensical - as this article makes clear.
"what does the keyboard sit on?"
Your hip, of course.
I have a wireless corded keyboard/tracball that's designed to be worn on the belt.
Yes, I can use it with both Linux and the BSDs. It was fairly easy to get used to ... easier than the switch to Dvorak, for example.
No, I don't use it. It's not conducive to the way I normally use a computer. The granddaughter tried it for a week, and then declared it "stupid".
Bought on a whim, which is rare for me.
ChromeOS is a desktop Linux with the Linuxiness stripped out. No choice about partitioning. No weird dual-boot mechanisms. No choice of desktops or package managers. No package manager!
Mint: I buy a new laptop. Spend 15 minutes installing Mint, all defaults, disable Caps-Lock, I'm done.
Been doing this for ten or twelve years with no issues, no real change. It just works.
My wife's Apple on the other hand constatly does inexplicable things. On the odd occasion when I am forced to boot into Windows (yes Adobe, I'm talking to you.) it's dear god, what a mess.
All of which is to say, why do Linux writers take such pride in the Grey-beard scenario? It benefits no-one, and surely does not benefit Linux.
> why do Linux writers take such pride in the Grey-beard scenario
Ah, isn't that paragraph you quoted - and other chunks of the article - Liam sarcastically ripping on that same "if it isn't hard it isn't Linux, what do those youngsters know!" attitude?
Sarcasm? In The Register? Shocked, I tell you.
As both a Mint and a Mac user, I agree with you 100% about Mint. However, I never experience Macs doing inexplicable things, which makes me wonder if this is a case of PIWNIC.
Edit: Moments after posting, I receive an email from my 80 year old mother, saying that her Mac is doing inexplicable things (yet again) and needs help. I know for certain that is going to turn out to be PIMNIC.
It certainly has inexplicable defaults - you can't tab between inputs without enabling "accessibility". I guess the semi-regular corruption of the NVRAM such that my not-sanctioned-by-apple USB-C docking station stops sending anything to the monitor could be closed as inexplicably but I think it's easily explained by corporate greed.
Macs regularly do inexplicable things. I know this, because I have the temerity to develop for the platform.
The moment you leave the One True Path, it all falls apart.
That path also changes every year, and you cannot just keep using your old, perfectly working software to do the needful because Apple will break it.
If it is running the Linux kernel it is Linux, simple as that really and that's how the stats should be shown, it can be broken down further from there if wanted, e.g. smartphones, Chromebooks, tablets, Ubuntu, Debian, Red Hat, etc.
So many people nowadays do all of their home computing from a smartphone, they are no longer wanting to pay for a laptop or desktop as well. So the way we think about categorising devices is less relevant.
But then, who put these stats out there? Do they receive any pressure to downplay the fact that Linux is the most popular of all?
> ChromeOS is a desktop Linux with the Linuxiness stripped out.
And that last line is the kicker - it is technically a Linux kernel but for the 'THIS is the year of Linux on the Desktop!' (TitYoLotD) people it needs to be flying its Linux flag high. The whole point was that users would have an epiphany and would dump Windows (and then later MacOS) like the plague and move to the enlightened world of the non-commercial OS. That is absolutely not the case with ChromeOS. They're not running ChromeOS because it's Linux, they're running it because the laptop was cheap, and for the most part they have no idea it's even Linux.
Now I'm not one of those TitYoLotD people, so I'm fine with counting those as Linux desktops as long as you realize that your number now has zero ideological relevance any more. Linux on the Desktop is no longer an aspirational open source dream, it's just the cheapest way for a different giant amoral corporation to sell you its products.
I've installed Linux Mint on the computers of several family members and acquaintances, a few total computer noobs among them. They never complain about it, it just works in a similar fashion like Windows.
They can browse the internet, read email, read and edit Office documents and watch YouTube and Netflix. That's sufficient for most people. Tech savvy teens will probably figure out how to install games like TuxCart with just a few clicks.
I hardly ever update their software, once a year at most. It doesn't matter since its much more secure in any case.
I hereby propose that everyone install Linux Mint on their family members' computer and see if they complain.
Ditto that. Just treated myself to a brand new laptop. It came with Windows 10 but I didn't even let it go through the initial install/setup, just plugged in a stick with Mint on it, booted into the BIOS and installed Mint straight away... begone Windows! A few days later (due to my purchase) Amazon sent me a code for £10 discount for Microsoft 365 family. LOL. I have zero use or interest in anything Microsoft nowadays. I can do everything I need using Mint.
Its not just mint. My latest laptop just for giggles I tried Fedora and Debian, both installed without a glitch. The only thing not working is the fingerprint reader but I can live with that. I had to go to Guix to get something that doesn’t work out of the box, pretty much 100% due to their “Libre” kernel.
Wow, what a leap of interpretation that the OP did it without consent?
For family and friends I now only support them if its a Linux distro as I have had enough of Windows malware and the stupid tricks MS have pulled over the years. I can set up a fairly locked-down Linux box and come back 5 years later, even when its gone out of support/updates, and it still works without malware on it. With Windows and paid-for AV I still found some family members had the box infested within 3 months of use.
I have some Windows machines and VMs for software that needs it, and I don't push Linux on folks that need Widows for specific use-cases, but my life is too short and precious to fire-fight the support problems that come with that territory.
There's another hidden benefit, too. It's much tougher for "Brian from Microsoft Security" who cold calls your Gran from Mumbai with frighteners about a virus. Good luck using the scam script to talk her through installing TeamViewer on Linux Mint.
"Now please to press the Vindows key and the R key at the same time, please".
MeDearOldMum and GreatAunt have been happy Linux users for many years now. Since moving them from the world of Windows to Slackware, their support calls have fallen from several times per month for the bastard child of Redmond, to none (zero, zilch, nada, 0) for well over four years now on Slack.
Linux works perfectly well on user desktops, as long as the wetware of the installer understands the needs of the user.
Another example: I have Vet clinics here in the Bay Area that I moved from AI/X to a version of Slackware in 1999. I also had clinics who chose to stick move to Windows.
The Slackware side just keeps on trucking, with absolutely seamless updates and zero security theater. The systems never go down unless told to go down. The only calls I get from these folks is for hardware issues (PEE CEEs and attendant peripherals aren't exactly well known for their reliability, when measured in terms of many years in a high hair/fur environment). The only formal training they received was back in 1999. They have no on-site administrator, as there is no need.
The Windows side, which I no longer support, has been nothing but trouble, especially whenever Microsoft rolls a revision out the door, and when the malware du jour strolls past their defenses. These systems are down for what totals in the weeks every year, crash fairly regularly, and need constant hand-holding by an employee who does nothing but look after the system. The only reason I know what's going on with them is because the Vets call me fairly regularly, asking if I can't PLEASE come look at their computers. I decline. I don't do Windows anymore.
It's pretty funny, at rad-rounds the Slackware folks are all on one side of the room, and the Windows folks are on the other ... The Slackware people don't want to listen to the constant bitching about Windows from the other side of the room.
What is mind boggling is the Vets using Windows insist that they HAVE TO "because compatibility" ... despite how that is obviously bullshit, given some of their compadres have been happily using Linux for about 25 years now.
I hear you. I sneak into offices and install Mint onto any PC I can gain access to, then I laugh at the employees when they go back to their desks and they can't even use terminal commands to do basic tasks. I leave a card for a helpline, but when I answer I start halfway through the instructions and make fun of them for not understanding the 40 other commands I have not listed in my initial response. When they get frustrated i refer them to another helpline.....which is also me, except I'm much crueller and I keep telling them to 'put the lotion on it's skin' and threaten to hose them
My understanding (which may be shaky - see icon - especially after the third can) is that ChromeOS uses the Linux kernel, the GNU userland but have done their own customisation of Xorg to run what is essentially a Chrome browser kiosk as the graphical layer.
Disclaimer: I use Slackware (30th anniversary 16th July hint) so what do I know?
Do the BSD folks include MacOS/iOS? It's a lot closer to a traditional BSD than Chrome OS is to linux.
Depends which level you're talking about. The kernel never was BSD, the user space was but has mutated. Trying to be an admin running a macOS system in a mostly Unix network triggers murderous rage(*).
(*) Well, it does for me.
"Generally, no, because it's a Mach kernel, not a BSD one."
If you ever boot it verbosely to see the boot text messages you get "Copyright the Regents of the University of California" the copyright disclaimer for the FreeBSD that kicks off the system bootstrapping to get it to the eventual loading of the Mach kernel then the graphical desktop.
> If you ever boot it verbosely to see the boot text messages you get "Copyright the Regents of the University of California"
You're both right.
The macOS kernel is XNU. It *is* Mach, and Mach is the original microkernel.
The idea of UNIX-like microkernels is that the kernel is -- well -- micro, it does at little as possible: memory management, process scheduling, and nothing else. All the OS functionality is achieved by processes running in user space, communicating via some IPC mechanism, called "servers".
In order to achieve decent performance, rather than a whole herd* of servers implementing the functionality of a UNIX kernel as a few dozen interoperating processes, the NeXT developers ended up putting all that in a single monolithic "Unix server" module, derived from BSD code, and moving that module into the kernel.
So what this means is that the most famous microkernel OS in the industry isn't really a true microkernel after all: it's a microkernel but with a big chunk of the FreeBSD kernel grafted into it.
In actual fact there are at least two real microkernel OSes out there deployed in large numbers in production.
QNX is a true microkernel, and has been since it was developed in the late 1980s. It is not FOSS, though.
A guide to the internals is here:
Minix 3 is a true microkernel and runs inside every Intel Core i-Series CPU, on the system management processor. It has shipped hundreds of millions of units.
However, it's not a very complete OS -- e.g. no SMP, lots of stuff doesn't work fully -- and as its creator Dr Andy Tanenbaum has retired, the project is sadly stagnant.
The L4 family has produced a number of kernels used in industry embedded in network controllers and things, but they are not general purpose OSes.
Plan 9 successfully side-stepped the entire question by moving the IPC into an existing well-known mechanism: the filesystem. Most of Plan 9 is user-space server processes, with a relatively tiny kernel, but it isn't trying to be a microkernel; they sidestepped the whole issue. The server processes don't talk directly to each other, they manipulate the filesystem. Everything really _is_ a file, and there are namespaces to isolate stuff.
P.S. AmigaOS is not a microkernel. Yes, the kernel is tiny, but there is no memory management to speak of, so there is no problem with making IPC performant: everything runs in a single share memory space and processes can communicate by just writing into, or reading from, each other's memory.
That is cheating and the costs were terrible:
1. AmigaOS was never very stable and every Amiga owner remembers "guru meditations" all the time when (not "if") it crashed.
2. They couldn't fix it. The single shared memory space meant the OS could not adapt to take advantage of MMU hardware when that got cheap and plentiful. The 68030 and later have an MMU built in, but AmigaOS can't use it, because it would break 100% of existing code.
* Yes, this is a gag. It's a cheap shot at a relatively _unsuccessful_ microkernel.
AmigaOS is an example of an Exokernel before MIT came up with the idea so it can't be that bad a design.
As for fixing AmigaOS, it needed more money than Commodore were prepared to invest in R&D. By 2.04 and definitely 3.0, documentation should have been updated and library calls which needed shared memory should have been complaining "you set the wrong memory allocation flags" if the programmer had done that.
Oddly enough, I would say generally yes, we do. Mach only replaces a small bit of the kernel, the rest is BSD. Source code is on GitHub and if you take a look then the usual BSD kernel files are there, and include the Berkley copyright in them, below the Apple bits. I haven't written kernel level stuff for MacOS, but I did for OpenStep, and you could simply take BSD kernel code and port it in a direct manner.
So, yes it's BSD underneath, and BSD at the command line too [though getting a little less so - vi to replaced with vim recently for example]. Sys-admining it, however, is *not* BSD-like, as someone else has commented. Don't try and configure things by editing files in /etc or you will end up very exasperated. But at a user level, if you are familiar with BSD you will find MacOS a lot more familiar than Linux is.
[ I suspect this is a moot point though, as the quantity of people in the world who are really only used to Unix and have no significant experience with Linux are pretty small in this day and age. That's me, but then I am old. And stubborn ;) ]
Last I checked (which was about 2 minutes ago), Unix is a registered trademark of... The Open Group. So if you conform to IEEE 1003.1 and pay some cash, you too can be Unix.
(However, if you're doing gas chromatography, that's a totally different registered mark, and GNU is deffo not that!!!)
What you're referring to as Linux, is in fact, GNU/Linux, or as I've recently taken to calling it, GNU plus Linux. Linux is not an operating system unto itself, but rather another free component of a fully functioning GNU system made useful by the GNU corelibs, shell utilities and vital system components comprising a full OS as defined by POSIX. Many computer users run a modified version of the GNU system every day, without realizing it. Through a peculiar turn of events, the version of GNU which is widely used today is often called “Linux,” and many of its users are not aware that it is basically the GNU system, developed by the GNU Project. There really is a Linux, and these people are using it, but it is just a part of the system they use.
Linux is the kernel: the program in the system that allocates the machine's resources to the other programs that you run. The kernel is an essential part of an operating system, but useless by itself; it can only function in the context of a complete operating system. Linux is normally used in combination with the GNU operating system: the whole system is basically GNU with Linux added, or GNU/Linux. All the so-called “Linux” distributions are really distributions of GNU/Linux.
And the last time I looked the typical GNU/Linux system had more lines of code from FreeBSD than from GNU. But whatever.
[I had real problems deciding whether I wanted the troll, fire or nuclear explosion icon for this. Readers should imagine all three apply. Yes, I'm a bit bored.]
Linux is a kernel. GNU is not an operating system as such either - unless you can find a GNU kernel to complete it. The GNU Hurd is available (I have vague recollections of another but I may be wrong there). Moreover if you include things like X.org graphical desktops, BSD-licensed Ethernet drivers and the like in your understanding of an OS, then any functional Linux-based OS is a little bundle of one thing and another, released under a rag-bag of licenses which are thankfully compatible under the hood. Distros also usually bundle in distinctly userland packages and branding, which may be less compatibly licensed. So a distro is almost never just "an OS" either. But we call it that because life's too short and Microsoft and Apple sell theirs that way too.
I've been using a Windows user-agent string for almost a decade, despite the fact I can't even dual-boot Windows.
A lot of sites absolutely lose their shit when they see a non-Windows/non-Apple UA, including my credit union and a couple of my credit cards. I MUST USE AN APPROVED BROWSER even if I'm already using Firefox/Chrome. I've even seen them assume I'm using a mobile browser.
At least I closed my account at the bank (Chase) that was setting .ru and .cn cookies.
I even have to keep updating the version number, as they also tell me to fuck right off if I'm not using the browser release that came out that morning.
A lot of sites absolutely lose their shit when they see a non-Windows/non-Apple UA
<voice mode="bewildered">Umm, err, but it's 2023!</voice>
Would I be right in thinking you're talking about US institutions? They've always seemed to me to be so far behind the curve that they're still in Flatland. About a decade or two ago I had a Bank of America investment account. They gave me online access to my (not insignificant) portfolio, protected by what they considered to be the highest password security. My password had to have all of a lower case letter, an upper case letter and a digit. It couldn't contain control characters or punctuation marks. It could be a whole six (count them) characters long.
Unfortunately NextCloud user forum sits on it. I haven't logged in for ages but if I click on one of the links in the occasional email feeds it gets arsey unless I use one of their choice of browsers. Unfortunately NextCloud itself seems to be going that way as although it doesn't complain it just doesn't present an login prompt.
I would say the core BSDs (Free/Net/Open) aren't rival sects, they're more different emphases. Free concentrates on running efficiently on mainstream hardware with lots of packages, Open focuses on security, Net wants to run on everything including toasters, bionic artificial limbs and engine management systems. Dragonfly is/was schismatic from Free.
minix & xv6 were pedagogic (as was/is Xinu which didn't get mentioned), so a different game.
Differently different, HURD and L4 are about microkernels, which arguably Are Not Unix Anyway.
Plan9 and Inferno are by definition Not Unix, being intended as research into what comes after Unix.
x7/86 seems to be an exercise in pure nostalgia.
So the only real division is between those Unices with Bell Labs ancestral DNA, and Unix-alikes. Even then, there's been so much cross fertilization that the distinction is very blurred.
On my SOHO network all the servers are FreeBSD (as is the OPNsense firewall/router) , the desktops are 1 each FreeBSD, macOS(*) & Linux, the laptops are 1 FreeBSD, 2 Linux & 1 ChromeOS, and the phones and tablets are all various Androids. No burning at the stake in this house.
(*) Or however it's capiTALisED this month.
and the Linuxi still can't understand why Windows is on more PCs than ever. I've tried a whole variety of Unix/Linux types/versions over the years and they are just as difficult to use/admin/maintain as any Windows system, and are still a minority system (on PCs) after all these decades. I'm no lover of Microsoft or Google but they work and for the vast percentage of people are good enough. I'm happy that there are Linux based systems for those who don't want to use Windows and one day I may be one of them, but currently I can't think of a single reason for me to switch. I'm sure that some of you are now starting to tell me how wrong/deluded I am and that I should switch to <insert distro here> but those are YOUR reasons, not mine. I use Linux every single day on my Android devices but they are a (mostly) user friendly version of Linux suited to the phone/tablet format. The closest Linux has gotten to my PC recently was a few VMs running under VirtualBox and whilst they were all usable, decent looking OSes, they never inspired me to replace Windows on my primary PCs. Any bets on the number of downvotes below? <LOL>
This post has been deleted by its author
I wonder how my personal contribution to the stats would look.
I use on desktop with the web/internet:
Windows 10 and Windows 11 on Intel
Mac OS on ARM (M2) and Windows on ARM (parallels) Debian on ARM (also parallels)
Raspberry Pi OS linux on ARM
iOS on iPad.
And numerous other server OS that don't ever reach the internet.
I have no particular enmity or preference for any of them. They all do their job and each one gets used for the purpose it suits. None of them are the best at everything. None of them give any problems.
If you're happy using Windows keep using it, the beauty of FOSS is that is and will be there when and if you need it.
There are not that many Linuxes, it is the same OS underneath, I like to compare it to beer and crisps, there are hundreds of brands, they all pour in a cup, come in bags, you drink them beers and chew them crisps all the same way. One likes some more than others and stick to one or two brands, every now and then you find a new brand that has some extra twist, but you mostly stick to the brand you like.
In my case I made the jump 12 years ago and I find myself in the opposite situation, I can't think on any reason to run Windows but I can think many not to.
I will tell you something though, a VM doesn't make modern Linux justice, bare metal is bare metal. Linux running on (supported) modern bare metal is great.
but 11 is a Fisher-Price like toy. Dumbed down far too much and full of adverts and bloatware.
My Windows 11 is a corporate image that is hardened and remotely managed by the admin. It has all the stuff needed for remote working. There are no adverts and no unnecessary stuff.
Of the three or four hundred Linux distributions known to Distrowatch - 200 odd are minor derivatives of Debian or Ubuntu. It's not a million Linuxes, it's about five. Red Hat is definitely an also-ran with only Slackware being much older than Debian - there's a month between them.
"and the Linuxi still can't understand why Windows is on more PCs than ever."
Oh, I understand that. It's a whole lot of commercial strong-arming of H/W vendors. What I still can't understand is why it's tolerated.
Having Windows and Linux on the same H/W I know from experience that Linux Just Works and Windows Only Just Works - somewhat reluctantly and very lethargically. At best Windows is slow, possibly because something is chewing up disk usage (even when nothing is apparently running the disk drive is in constant use) and when it comes to updates -- Windows folks, it doesn't have to be like that.
"Having Windows and Linux on the same H/W I know from experience that Linux Just Works and Windows Only Just Works"
But I don't use Linux, or Windows - I use the programs that run on them - and Windows wins that hands down. You can talk about Libre this and thunderbird that, but with a system that's fragmented between even 'just' 5 main distros, developing commercial software and reaching the fragmented markets for it is a hassle; and that's after the battle has been lost - it was a real mess 25 years ago when there was still a chance.
What Google did with a new market was provide both the OS and a market place for one consistent platform (okay, one that evolved rapidly, but only that one). We ended up with a Chrome / IoS duopoly precisely because markets only ever seem to handle about 2 competitors - meanwhile Linux is busy competing with itself in 5 (or 205, depending) markets, to the detriment of all its parties.
Since Chrome OS is basically a heavily modded Gentoo Linux, I'm fine with that operating system (along with Chromium OS and CloudReady OS) being counted as Linux.
However, the finished product Android as seen on phones and tablets is more of a de facto proprietary operating system so I would not count it as a standard Linux operating system although it is a Linux kernel-derived operating system.
Wannacry was the result of the NSA hoarding exploits for closed-source software instead of responsible disclosure. I strongly suspect if it was open source someone more adult would have discovered the flaw first and patching on Linux is for the most part an absolute breeze, unlike Winwoes.
Anyone could have found that vulnerability in Windows, but not enough people looked hard enough to do so and not enough people patched. Linux isn't immune to vulnerabilities that hang around. Various commonly-used packages have been subject to bad bugs that have been there for years or even decades. People also don't always immediately patch, hence why I had some fun finding boxes that were still vulnerable to Shell Shock months after that became well-known (for context, that bug was in code from 1989 to 2014). The only question is how many people will exploit something when it is discovered and how quickly they will do so. Millions tried it with Shell Shock, but if every consumer computer was running Linux, that would have been even more people. Linux does not provide you a security guarantee, and if you act like it will, you open yourself to risks that you don't need to.
Had there been one single Linux distro that could have been true. People don't chose choose a computer for the technical merits of the O/S, they choose it for the applications (& games) they need to run on it. When an application developer decides where to focus effort they are firstly going to go for Windows and then bitch when new versions of Windows come out. Were they to also produce a Linux version they would first need to decide which Linux variant to go for. If they had been just one Linux variant to select this would have been easy but with dozens to choose between it's easier and much less expensive to just to ignore them.
The only significant Linux variant is systemd and it is a moving target kinda like Windows in a lot of ways. GNU/Linux without systemd requirements can be a single target for a develpoer. Targeting non-systemd GNU/Linux should also work on a systemd GNU/Linux until it diverges too far (not if, but when). Granted, there are different package systems but there are also for Windows (I supported Windows from v3.0 to reluctantly current.) There is not that much difference in the package systems, I know this because I have manually extracted from .debs and .rpms to install software on my machine when a slackbuild does not exist. I have even installed supposed systemd required software without systemd, it only looks for systemd indicators or uses systemd service system (aka daemons outside of systemd). Vast amounts of Windows software also erroneously claims the need for "admin rights" but in reality does not. Developer laziness is a major factor in all of this.
I can take an application's binaries compiled on my system, for my system (Slackware 15.0) and install it on a systemd Mint installation, provided I compile with the same kernel version and compatible library versions.
Windows application binaries have a somewhat smaller target of kernel versions and library versions but it is still a factor. Lazy developers had coded windows version checks for NT based Windows to halt on "Windows 9"x systems originally meaning Windows 95, 98, and ME. This was a major factor in Windows jumping from version 8 to 10 and skipping 9,
Just check and my current Linux Mint Firefox browser reports as being
"Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:109.0) Gecko/20100101 Firefox/113.0"
Which is the user agent I set to avoid badly written websites moaning about me visiting using a computer that wasn't Windows or MacOS. So my browser stats wouldn't be counted in their stats.
The only downside I see to this is that many dev teams use these stats to decide what to support. If it looks like more folk are using Windows/Chrome than they actually are, then the project managers out there will have more reason to drop support for *nix/Firefox. This probably applies more to browser than OS, but worth calling out all the same.
Came to the comments to point out this very thing - a lot of Linux users are likely reporting a false user agent so websites won't complain. In other words, the testing methodology mentioned in the article is fundamentally flawed; the real number is likely much higher. (How the routine blocking of 3rd-party traffic, which I suspect is how Statcounter gets its data, would affect the results is anyone's guess.)
(I only use a fake user agent string on sites that I really want/need to visit but don't work properly without it.)
Ok, so typo, UNIX, but then that will cause queries.
Linux is just the setting free of an idea that started before that.
I fell in love with the beauty of the IO model and the 'everything is a file' thing.
My degree was in OS internals and compiler writing, and UNIX (ish) provided the simplicity and scalability that made it a beautiful foundation.
I have read a lot of comments here, and the more i think about it, yes chomeos is Linux, but it is also more browser than linux. So they could change the OS to something different and it would not change what a chrome book bis to most users
I'm a retired Android developer and a big fan of ChromeOS Flex. It allows me to reuse the laptops my kids used at uni and run Android Studio on very modest hardware. It's much more responsive than any version of Linux I've ever tried and gives me exactly what I want: a browser, a media player and a file manager. I also make full use of the Google suite of apps and am amazed that people still store their data locally. When I do run Linux apps, the tight integration with the file manager and the text editor is much appreciated, and the diagnostics app is a hidden gem.
My intuition tells me that ChromeOS does not run a full Linux kernel, but only those services that it needs. That is what I would do myself and it would explain how it is able to outperform lightweight distros such as Atom.
amazed that people still store their data locally
Just lop "locally" off that phrase to get to the reality. Either you store your data or someone else stores it. What amazes me is that people trust someone else to store their data for them, not as a backup* but as the only store.
* There's scope to have concerns about that as well.
> "...think ... is junk, ....think everything else is junk, ....folk think they are the true cutting edge, ....consider everyone else newbies, ....folk think all the rest are still in the stone age somewhere..."
I'm inclined to agree.
> A MILLION different Linuxes,... and the Linuxi still can't understand why Windows is on more PCs than ever. I've tried a whole variety of Unix/Linux types/versions over the years and....
I've just binged my biennial ISO-romp, loading several distos into a virtual or real machine.
As a >40-year geek (ADM3a), this stuff is TOO geeky for me. Or maybe just un-finished? Not ready for 99.8% of users?
Of four well-known distros in this round:
One did NO memory check before installer lock-up, twice. (I had left the VM default RAM; but any dang bootloader should check RAM ASAP and complain, basic self-defense; this one could not be bothered.)
Three others got to a desktop with NO workable menu system. Yes, I know how to use a console (if I can find it) and type commands. And blind-guessing 'firefox' did launch the browser. But people I know, only my father would do that (he trained RCA's card-base JCL). If I want a "bare" screen, let me sweep the cruft off; don't leave me click-less from the get-go. Win3.0 and old Mac did this better.
If I did get a menu it was unhelpful. Who makes-up these names?? If I only knew Win/Mac, should I know what 'Evolution' (a dirty word in some places) is? A lot of lesser(?) app names look like leftovers from a Scrabble orgy.
If I did get a menu, twice any slight mis-click made it vanish.
If I did get a menu, it seemed un-curated. Some had less than a dozen, half uninteresting. One turned out (after configuration) to have thousands of apps, 99.9% not user-friendly. For all its new and legacy faults, MS does not put a disk partition tool at the same menu level as an office suite. (I understand the commercial reason, but recognize that non-hobbyist machines need W0RD thousands of times more often than fdisk.)
How many "Settings" apps can one OS need? (Yes, Windows is open to this complaint.)
Two had MOST dialog boxes bigger than a small LCD or a VM window. The OK button way below the bottom of the window. Just like RAM, there seems to be a 'wot th hell' attitude about resource checking.
What is with the UGLY wallpapers?? And desktop pickers without any "NO Image!" choice? Even Win3-10 has no-image choice; on small hardware it can be snappier, also less distracting. (The blurbs describe "inspirational" desktops, but if you are WORKING how can you see it?)
(Do not get me started on transparent windows.............)
Partial exceptions: MX linux 21 and (coff) Ubuntu Mate
Downvotes start below.
" ... and the Linuxi still can't understand why Windows is on more PCs than ever. ... "
No, actually many of us know EXACTLY why it is. Only those that haven't been around awhile, or have no use for history can't understand it.
There are basically two main reasons why Windows is on "more" desktops.
The first is because in the beginning Microsoft said " Our OS and nothing else. Or no Windows at all. Fuck you." And they got slapped for it.
Secondly, it was and is, technically, "You can't sell a computer without an OS. If you do, no Windows for you. Fuck you."
So yea we know exactly how and why Windows was and is still the dominate desktop OS.
And being the "best, most secure desktop OS" doesn't have a fucking thing to do with why it is. Because subjectively, it isn't.
Downvotes for truth, apparently!
That linux distros so rarely start up (first boot) with anything more than 6x800, yet use menus designed for 1024+ infuriated me 20 years ago, and apparently it is still true!
(I know it is, from VM installs)
Bloody Stupid Johnson lives!
Downvotes because many of us have installed Linux a lot of times and know that the only way that can happen is if the display has had to fall back to some very basic driver few have us have ever seen and there are few circumstances which can cause that. One is that it's being installed in a VM which isn't offering a better functioning emulated graphics device; that says more about the VM or the abilities of whoever configured it than about Linux. Another is that you've chosen a really hair-shirt distro that contains few drivers and not one for your H/W or maybe one where you actually have to nominate the driver you want and have selected none or the wrong one - just use a mainstream distro. A third is that you have a bleeding edge graphics device for which there are, as yet, no available drivers.
With any of the mainstream distros you have to put some real effort into dropping the installed into a mess it struggles with. The OP told us how he did that - with a VM.
From what you said, maybe you haven't tried installing Linux in the last 20 years.
I just had the opposite experience. I had to reinstall windows 11 on a friends laptop. In fact he wanted it setup as a dual boot with Ubuntu. For windows I popped in a flash drive for the install and fired it up. Got to the partioning screen - there are no hard drives showing up because it needed intel rapid storage drivers to recognise the hard drive. Of course, Intel no longer upply these as a zip file, only exe which isn't much use when the only windows machine available is the borked one. Once I eventuallyfound the drivers I restarted the install, to be fair it was simple enough, but there was no mouse pointer, so it was keyboard all the way.
Even when it booted into the 'welcome experience' where I normally tell it not to spy on the user as much the screen was low resolution and there was still no working trackpad. I had to navigate past the part where it was connecting to a network using the keyboard so that it could download the correct drivers. Even then I'd finished the whole config and waited about 15 minutes before the mouse pointer and screen set themselves up properly. Then there were the updates. Windows update kept failing at installing updates, because it had already sucessfully installed the updates but for some reason kept thining it didn't. Apps on the MS store wouldn't install initially when it was trying to restore those, I had to go in cancel the restore and manually click install for several of them. For the average user all of that isn'tsimple stuff. My friend had had a go before but gave up and called me.
For the Ubuntu install it was all fine, usb flash drive inserted, hard drives recognised, partioning was as simple as telling it I wanted to dual boot. The screen resolution was right from the get go and I had a working trackpad. A couple of minutes after install I was prompted to reboot and updates just worked. Much easier than Windows for the average user of the system I was using
Keeping this comment kind of on track, I had a play with ChromeOS flex a month or so ago, the install process was on par with Ubuntu bar the dual boot stuff. However with the linux stuff enabled there were a few quirks and issues - particularly in rendering firefox menus and file mangement between the chrome bits and the linux bits. I'm sure I'd have gotten used to it and based on my experience with it I would definitly class it as Linux.
You gave us a clear hint as to why things are going wrong - you're using a VM so the installer is having to put up with whatever unrealistic settings you've provided. Free up some disk space and let the installer install into that.* The it will have the real disk, the real display H/W and the real memory rather than whatever simulations and short measures the VM you configured provides.
* It's possible that some installers won't give you that choice and want to use the whole disk- if so, back away from them. I haven't looked for years, but does Windows still fall into that category? It used to be that if you wanted a dual or more boot Windows and anything else you had to install Windows first, shrink the partition and then install whatever other OSes afterwards.
So you don't actually understand what the headline or article is saying?
Think about it. The article is make the case that ChromeOS IS in fact a Linux distro. So if you count it's 4+% of the market, then LINUX has a 7% share. So both the headline and the article are correct. Linux, as most people think of it, actually is half of the Linux OS market.
I'm sorry if it's a bit complex or convoluted, but it is in fact correct.
There is a kernel called Linux. A kernel is not an operating system. Application support is what makes an operating system. In fact, this is reinforced by the various Linux system call interface shims that have existed for Windows (specifically, WSL-1), FreeBSD, BlackBerry 10, Solaris which permit Linux-compiled binaries to run without a trace of the Linux kernel being present and with no virtualisation.
If by "Linux" one is going to refer to the various piles of userland applications one gets in the average distro, one has also to include operating systems like Windows (WSL-1), FreeBSD, BlackBerry 10 and Solaris as being "Linux" too, because those userland applications can be run on those OSes too. In which case, you can say that Linux has about 80% of the desktop market (or whatever share it is that Apple hasn't got). Restricting the definition to those OSes that happen to use the Linux kernel is nonesense.
(Let's see how that flies!).
I do think application compatibility is part of the picture, but I'll point out that most of those shims only work with some kinds of applications. For example, try running a Linux-based service using one of those. Most of those layers have left out the services stack, because it's not an operating system, and haven't plugged in Linux services into the host's service stack either. As such, they're similar in spirit to virtualization, because they have an environment in which you can run other things which is not the same as the host. Of course, the definitions get blurry here, because there are multiple service systems in Linux kernel OSes, so there's not a guarantee of compatibility there either. Still, that's why I don't consider Android to be a Linux system; their applications are incompatible with each other in nearly all cases.
>As such, they're similar in spirit to virtualization, because they have an environment in which you can run other things which is not the same as the host.
Not really; if the Linux system call interface is supported, then it is native to the host. The fact that the host can also talk a different system call interface is also native behaviour. It's only "tradition" and "branding" that makes one of them senior / more important in that operating system.
WSL-1 is especially native. It's the only system call interface publicly defined on Windows. The other is masked by Win32 which is a DLL that knows how to call the Windows NT kernel interface. The actual Windows NT kernel interface is not available to developers, no one knows the details officially outside of Microsoft. Whereas the WSL-1 Linux system interface is genuinely a kernel system call interface, in that you can load a set of registers and raise the right interrupt (or whatever) and the kernel steps in and does its stuff. Sure, most applications would access it through glibc, but there is nothing stopping use of alternative libraries, or just writing your own to make kernel call directly.
One can argue that WSL-1 made Windows more like a Linux than "Windows", in that it has to now honour a system call interface that is defined and controlled by the Linux kernel project and ultimately Linus Torvalds. If Linus makes a change to that interface (something he seems sensibly and supremely unwilling to do without very good reason), MS is under pressure to follow suit. The power to change a thing is to have control over it, to own it in a way; Linus has a control over Windows now. Whereas MS can make changes to the NT kernel system call interface all they want, thus breaking compatibility with previous versions of the NT kernel, raising the question "What is Windows, really?". Sure they can update Win32.dll thus restoring compatibility and no one is the wiser, but things have changed radically under the hood.
There's a reason why we now have WSL2; this uses hypervisor technology to run Linux in a micro-VM. This increases compatibility (eg "docker" won't run on WSL1, it runs fine on WSL2). But at a cost. For example, with WSL1 your processes show up in the windows task manager but with WSL2 this doesn't happen; Windows is mostly blind to what happens inside the VM.
Microsoft can easily deprecate WSL1 if they need to, but retain the ability to run Linux apps via WSL2.
Yes I know that MS chose to "finish" their embracement of Linux by not completing WSL-1 but by integrating a VM instead. It works well, I use it frequently. But this is kind of orthogonal to the matter of whether Windows is also a Linux by virtue of WSL-1 supporting very large swathes of the system call interface. By some measures, it's more Linuxy than Android!
WSL-1 Hides a Massive Change?
In implementing the Linux kernel's system call interface, there is a clear sign of a large internal change within the Windows kernel itself. This is related to the select() function, and its equivalents epoll(), etc. Linux, where everything is a file descriptor, and every file descriptor more or less behaves the same (i.e. very *nix) has the system calls required for glibc to efficiently implement select(), epoll(), etc. Ultimately, it all boils down to devices raising interrupts, service routines handling them, the kernel working out what's what, and returning control back to the process that is blocked in select() with an answer it can act on. Meanwhile, the process has been suspended pending a relevant interrupt being raised, taken no processing time, etc. All very modern *nix.
Windows, though, only ever implemented select() (or any of its equivalents) for IP sockets, and nothing else. There is no way you can in Windows have anything like select() operating on (say) a pipe, or a serial interface, telling you that the stream has become ready to read. It's IP sockets, or nothing.
You see the consequences of this in various places. ZeroMQ does not support an ipc transport on Windows, because you cannot achieve an implementation of select() that works on Windows pipes. Cygwin, when that was being implemented, the team came to the select() function call and found there was no way of having a complete implementation either. There's some hilarious chat on some very old dev chat lists about it, where it dawns on them that a proper, interrupt driven select() couldn't be made to work on Windows. Their solution was to start up a thread per Cygwin fd and poll the underlying device handle. Very inefficient, very old *nix (like, dark-ages early *nixes when the idea of a *nix was quite new, and threads were a non-existent concept). Boost's network library is all about registring callbacks rather than waiting for input, becase Boost realised that that was the only way to get Boost working fully on Windows.
The reason for all this is that Windows (and its kernel) is fundamentally a proactor - you register a callback to handle a handle data that has already been read - whilst modern *nix is a reactor. You can implement a proactor architecture easily and efficiently on a reactor (e.g. Boost's network code for *nix), but it is impossible to implement a true reactor on a proactor (hence Cygwin's multi-thread polling bodge, and ZeroMQ's complete lack of interest in ipc on Windows).
However, in doing WSL-1, Microsoft had to support the select() function in all its glory as implemented in glibc. Write code for WSL-1 that uses select() on sockets, serial devices, pipes, etc, and it works. The very, very big question is, "How?". Possibly they did the same as Cygwin, but inside the NT kernel a no one is the wiser. Or, they worked a significant overhaul of the NT kernel so that interrupts on devices could be used to wake up a process stuck in a select() call inside a WSL-1 process.
If it were the latter, then in principal Windows applications could also be given an API call that allowed for reactor / select() type functionality on more than just sockets. But, so far as I'm aware, they've not done that. Which I think is a great pity, because the lack of full reactor support on Windows is one of the things that makes it truly difficult to port applications and libraries.
If someone said "Android is not a Linux," that is defensible. You can't easily download it for free, and you can't run it on your own generic off-the-shelf computer.I've run it in a VM on my Kubuntu systems.
On its native platform of smartphones and tablets, you can't run ordinary Linux apps on AndroidI also have a command-terminal text-editor (microEmacs) which I build for and run on my Motorola phone. OK - it doesn't display to the phone screen - I have to ssh into it and run it in a Konsole terminal - but it is running on an Android installation (which is the only reason I build it - because I can and it works).
" ... I wonder if most Linux users actually feel this way or if the people who think ChromeOS isn't Linux are just the loudest. ... "
My money is on that most Linux ( okay, maybe not most, but fucking far to many ) don't actually have a fucking clue as to what Linux actually is.
Linus is the kernel.
Ubuntu, Redhat, Debian, ect. are all OS's that use the Linux kernel.
I use opensuse OS. It uses the Linux kernel.
Most people running a Linux kernel have no idea they're doing that, or care about it in anyway, nor would they notice if, say, Google swapped out the kernel for something else (like Fuschia).
The one thing Steve Jobs got right was, "It's the software that matters". And there's an awful lot of "software" and associated services these days written to not care at all what the OS is, or care what web browser it's running in. Apple themselves do their best to ensure that software on their platform is specific to it and non-portable...
"My money is on that most Linux ( okay, maybe not most, but fucking far to many ) don't actually have a fucking clue as to what Linux actually is."
Of course not. A lot are running an Android userland. Others aren't aware of using anything because it's tucked away inside their router, or their TV. As you point out Linux is the kernal.
"Linus is the kernel."
No, Linus is the original developer and continuing head honcho. These days I'm inclined to pass on typos because I make too many myself but I'll make an exception as you seem not to have learned that it's possible to make a point more clearly and more forcefully without expletives.
"... typical Linuxes are tools for nerdy hacker types"? I for one strongly disagree.
I have used many flavours of Linux since the mid-90s. For many years now Mint and LMDE are my number one (and debian as host for VirtualBox).
End users get Mint from me. And guess what? Even the most non-nerdy ones are happy with it. Mint is end user friendly to such an extent that it can beat Windows.
My stance is: Anyone who can read and write can use Linux.
The ones I am talking about would never ever do a search (of course not goggle) and manipulate the system on their own.
After I have installed the system, it just runs. And runs and runs and runs.
Only when an upgrade is necessary (not earlier than four years) I enter the scene.
While in the meantime the process of upgrading is automated and works so seamless that any reasonably intelligent person can do that.
Android is a Linux distribution. It's just not _desktop_ Linux.
It has a native shell that you can reach without needing to root/break/hack; just enable debug mode and "adb shell" (just like you need to put ChromeOS into developer mode to enable crosh shell). Or you can install termux app (from the app store, so easier than enabling crosh) if you want something more fully featured.
No, it doesn't come with glibc, but then neither do other linux distros like "Alpine Linux", nor other embedded Linux systems (eg OpenWRT).
Because it's not a desktop Linux, programs written expecting a desktop won't easily work nor be portable. But CLI programs and the like mostly just work as expected; you just need to compile for the target hardware and libraries.
In every respect it's Linux; it's just not _desktop_ Linux.
> In every respect it's Linux; it's just not _desktop_ Linux.
It's a valid take, and one a few people have put to me on Twitter and by email.
I think it does demonstrate the complexity of what superficially looks like a simple question.
My greater point is one often expressed in other contexts -- for instance, as Sinclair Lewis said in 1935:
"If fascism came to America it would come wrapped in the flag and whistling 'The Star-Spangled Banner'."
When something happens that many say will not happen, they will not recognise it as happening, because it will not be in the form that they expected. But it will happen nonetheless.
I taught a Linux-based university course on an Acer Chromebook with no issues when my 7 year old Dell laptop went nuts (during lecture!) early in the semester. It ran the same tools (slightly different versions) with the same presentation and results.
It is indeed just Linux with a default web-based desktop.
Like the other 95% of desktop users I've been using 'something else'.
Linux on the desktop is an almost vanishingly small minority not helped by all the stupid spats over which 'distro' is best. Nothing that happens in the 'Linux desktop arena' is of any importance beyond the handful of people still clinging to it.
Face it. As a desktop operating system Linux is an abject failure. Possibly not for technical reasons but still - it is providing a desktop experience to such a small percentage of users that really if it reverted to being nothing but a console-based server OS the world wouldn't care.
> MacOS could also be considered to be 'a Linux'
No, it couldn't. As discussed elsewhere in these comments, macOS runs on top of the XNU kernel, which is a combination of the Carnegie Mellon Mach microkernel, plus a large chunk of the FreeBSD kernel.
There is no Linux anywhere in macOS whatsoever.
> Then Windows must be Linux too because of WSL.
Nope. WSL 1 does not contain a Linux kernel at all; it is basically emulated by a Windows subsystem.
WSL 2 does container full Linux kernel, running inside a VM. No containers here. The operating system colonel in control of the hardware is the windows Colonel, not the Linux kernel.
So, while effectively WSL2 is a program running on top of Windows, you could also argue that by any reasonable definition WSL2 *itself* is a Linux... that doesn't mean that the parent operating system is one as well.
...is all the stupid bloody names they use for parts of the system that could just be named something sensible. For example:
Windows: File Explorer
Linux: Nemo, Nautilus, Krusader, Thunar, Dolphin, Kaja...
It's just *confusing* for the average user. And these things change all the time.
And then there's the browsers. There's no easy way to just install 'chrome' - yes, you can install Chromium. And yes, there's firefox. But people like to use Chrome and are familiar with it.
What I'm talking about is the ease of use for ordinary home users or small businesses. People want compatibility with everything else and they want simplicity. Moan all you like about Windows if you're a Mac user and vice-versa but at least it's relatively simple to get stuff working with both those systems. Why would any small business try linux when:
- it won't run Sage
- it won't run Chrome (yes, I know it runs Chromium but not out of the box)
- it doesn't come with proprietary drivers (the average business user couldn't give two hoots about 'but it's not open source' - they just want to get things done!)
- it doesn't run MS Office (yes, I know it runs LibreOffice but I've tried introducing that and there's a reason that companies will use paid-for-MS365 over free-libreoffice, and that's because it works with everything else that they, their suppliers and their customers use!
- Cloud file systems: only dropbox has a linux client that's relatively easy to install.
It mostly boils down to lack of familiar software and a relatively poor out-of-the-box experience.
Obviously there are use-cases where a linux desktop would work for a small business, but what's the advantage when the computers cost practically the same and there's no guarantee that the services you use will continue to work over the years?
I know I'll get downvoted for this, but public opinion does seem to support these arguments. If it was really that good, and that cheap, then why aren't more people using it? You can't have good, cheap AND easy all at once.
> There's no easy way to just install 'chrome'
This is wrong.
They're absolutely is: what are you do is you go to google.com/chrome, press "download", download the .deb or .rpm package, then you double click on it in your file manager.
You say yes a couple of times wait a few seconds and then you are running Google Chrome.
Alternatively, of course, you could go to an App Store, such as Snapcraft or Flathub, type in chrome in the search box, and install the first thing that shows up.
So, no, your claim is incorrect in every important way. It is about as easy to install Google Chrome on any general-purpose Linux distro as it is on Windows or macOS.
It is easier and quicker to install a general purpose off-the-shelf distro such as Mint than it is to install Windows itself, then patch it, updat it, install all the various drivers it needs, install a whole suite of applications, configure the desktop, clean up the start menu and so on.
That is about a days work for a standalone machine. Any competent reasonable sized organisation will have a ready to use image that they can push out onto a new client machine but that is not usually an option for a very small business or a sole trader.
I recall when Chrome OS first appeared there was a lot of talk among Linux fanbois along the lines of "it's not a real OS" and "there will be no uptake" and of course "it will never last". Of course it's not up to those fanbois to define what is and isn't a "real" OS. However the big problem for those fanbois is that now that Chrome OS is not only the most popular desktop Linux OS but it has a bigger installed base than all other Linux desktop distros put together they can't suddenly accept it into the fold without losing face.
That being said I think that the problem with Linux and privacy conscious users i not the kernel "per sé" (yes, it needs that accent), but Google's total control of Android and Chrome and Chrome OS. I don't trust Google, simple as that. My Lenovo Chromebook is a wonderful and very light travelling companion. It just works, and that's great.
So does my HP Ryzen 7 with Sid (and MATE). It also weighs about 700 grams more than the Lenovo Chromebook. And it doesn't have a touchscreen.
TL;DR - I think it's all about the way you feel about privacy. And Google doesn't guarantee you that, it's its own business model that requires it. Debian, Slackware (my first distro) and many others do not profit from your data. That's the real difference. You might dislike my choice of MATE or any other DE, but that doesn't undermine the commitment to privacy of FOSS software (like Firefox).
Cheers fellas! :)
Google's Chrome OS certainly is based on Linux.
Back when Chromebooks first came out, though, comparing the possible major alternatives available - a Macintosh, a Windows PC, and a Chromebook - the Chromebook couldn't keep resident applications on a local disk, it could only run software it got off the Internet. That got me to reject it as not a computer, but a useless toy.
They've now fixed that limitation of the Chrome OS. But even so, I am strongly sympathetic to those who reject it out of hand. Maybe not being truly Linux isn't really what its problem is, but it's less flexible than Windows, while Linux is more flexible than Windows.
Since Chromebooks can't normally run typical Linux software (no package manager) then they don't increase the size of the market for typical Linux software. That is the most fundamental sense in which Chromebooks fail to enlarge the market share of "Linux", and I think this is a sufficiently real concern that taking notice of it isn't hair-splitting religious warfare.
It isn't what Chrome OS took from Linux that counts; it's what it gives back to Linux that matters.
Although at first it looks that Android is a completely different thing than any other Linux OS, that kernel is really just a hacked Linux kernel.
It is possible to run a userland on top of Android than is more in line with what is know as a Linux OS. Apps like UserLAnd and Termux will run a userland known from Linux without emulation on the Linux kernel of Android.
As some experiment (and just the question "Does this works?") I installed the UserLAnd Android app on ChromeOS, installed Wine and tried to run a Windows game. Yes, it did work..
Also, on ChromeOS the Android environment (ARC++) does run directly on the Linux kernel beneath.
So the question remains: what is exactly a Linux OS? :)
Hi fellas and galls. I’m a newbe. Just got Mint running on an old Toshiba and am happy as a clam. Here’s what I think. Most average people - like me - don’t give a crap what OS they use. They want to do a couple of basic things on their computers. Actually. Less than before. Mosty it’s email webpages archiving pictures and some light documents and spreadsheets. Let’s face it. Most of us are not programming stuff. And actually as smartphones do more and more I find the use case for a full fledged computer less and less. That is why many people are totally happy with chromebooks. I am not. Because google can suck a big one. And I’m not into windows because they are now a subscription service. So my only option when windows expired on my old machine was to buy a new windows or Mac or Chromebook. OR - I could put on mint and be on my way for $0.00.
So that’s it. No one cares what OS they use. They just want something that works.
But I agree. Chrome needs to do more to feed back into the Linux community. But it won’t. Because google can’t make money on that.
In the meantime. I’m happy with my mint machine. Tootles.
Personally I think the very idea of combining the numbers of OS's related to linux then state that as Linux use cases is as daft as suggesting of all the 22 species of apes including gorillas, orangutans, chimpanzees, bonobos, gibbons and humans, only humans are included in population counts even though they all originate from the same family tree as proved by their remarkably similar DNA, but just the wrong DNA.