Skip Navigation

Linux has nearly half of the desktop OS Linux market

www.theregister.com Linux has nearly half of the desktop OS Linux market

It would be twice as much if people could stop arguing about who counts and who doesn't

Linux has nearly half of the desktop OS Linux market
55

You're viewing a single thread.

55 comments
  • A quick summary of this [frankly, rather crappy] article boils down to "I don't understand, why people ignore the 4% of ChromeOS usage when talking about the 3% of Linux usage?" written in the most facepalm-worthy way, spamming fallacies like there was no tomorrow:

    • "Which is also Linux, but the wrong kind of Linux." - strawman
    • "which means that [desktop Linux]¹ has less than half of the [desktop Linux]² market." - ambiguity ("1" refers to a subset of "2")
    • "It's not a typical Linux, because typical Linuxes are tools for nerdy hacker types, and that kind of OS will never, ever go mainstream unless someone forces people to use it" - begging the question + ad hominem
    • "So naturally the Forces of FOSS hate it. Of course they do. And how do they express that contempt? By saying it's not a True Linux." - repeating the strawman again, for an ad nauseam

    Also note that the same type of stupid reasoning from the article would also "prove" that MacOS is a *BSD.

    What a weird hill to die on.

    • But nobody considers MacOS a BSD.

      I thought the author's point was pretty well-made. Some implementations are so far removed from the original's goals that they no longer count for it even though they're technically a match.

      It's merely an oddity that ChromeOS is based on Linux; it could be based on anything else and it wouldn't change anything. It doesn't work like Linux does and you can't use it for most of the things that make Linux worth using.

      Even Windows (via WSL) is more Linux at this point than ChromeOS is. Let that sink in for a moment and you'll see why we don't count ChromeOS.

      • Except the author is arguing that ChromeOS should count at least in desktop Linux statistics, he is criticizing the FOSS community for not doing so for ideological reasons more than technical ones.

      • But nobody considers MacOS a BSD.

        Yup - and that's the point of the comparison. They share a lot of important code (mostly in the userland IIRC), and yet people don't consider MacOS a BSD, since both drifted away so much that lumping them together is simply not useful. And yet it's what the author is doing with ChromeOS and Linux, that are in the exact same situation. (Another way to say this is how you did it - they don't work the same and you can't use one to do the things that makes the other worth using.)

        I thought the author’s point was pretty well-made.

        The author's point boils down to "actually you got 7% Linux, not 3%". I don't think that it was well-made; it relies on disregarding why people don't call ChromeOS "Linux".

        • MacOS is not a BSD specifically because its kernel is not BSD (although it has some bits of it), so the comparison isn’t really sound. What makes it “a BSD” (or Linux) shouldn’t be the graphical environment.

          Immutable distros are making accepted “Linux distributions” even more like ChromeOS, while ChromeOS has gotten more like those immutable distros.

          • Sorry beforehand for the wall of text.

            Even if FreeBSD and MacOS shared the exact same kernel, I don't think that people would say that MacOS is a *BSD - due to what lemmyvore said, it does not work like a *BSD. The code itself doesn't matter that much, and this is not just about ideology (more on that later); it's about how the software interacts with the user, the machine, and the rest of the world.

            I'll highlight this through a thought experiment. Imagine an OS called DuckOS.

            Let's say that you refactored its source code, compiled the result, and released the refactored version as NaN-OS. No DuckOS line of code was left untouched; you refactored everything. But both DuckOS and NaN-OS still behave exactly the same, they run in the same machines, you even reimplemented DuckOS's bugs and quirks.

            For all intents and purposes, NaN-OS is still DuckOS. It quacks like DuckOS, walks like DuckOS, then it's still DuckOS.

            Now let's look at ChromeOS vs. what people usually call Linux (the 3%):

            • ChromeOS: thin client OS with mostly web-based interface, running software from someone else's computer ("the cloud"). Your data is also in someone else's computer. You can access both from other ChromeOS devices. You aren't really expected to mess with the system itself. Mostly intended for secondary devices, that become fancy bricks without internet.
            • Linux (the 3%): new software is installed in the system itself, as usual. Your data is by default stored locally; you could store it in someone else's computer like the above, but that would be an atypical. Intended for generic purpose machines, from supercomputers to toasters. Customisability and adaptability are some of its appeals.

            Sure, ChromeOS runs the Linux kernel... and then what? It doesn't behave like that collective mess of Linux distros.

            Now, regarding ideology. In another comment, you mentioned that the author criticises the FOSS community for not calling ChromeOS "Linux" for ideological reasons. I do agree that ideology plays a role here, but it's neither a factor that should be brushed off, nor the sole factor. When the FOSS community says "use Leenooks!", they aren't saying "use the kernel" - you can bet that most would be happy campers if you used hurd instead. They're instead saying "use that bag of customisable and adaptable OSes, built upon open source in letter and in spirit". It's hard to claim that ChromeOS is part of that bag, even if it happens to use the Linux kernel.

            (And for the sake of statistics it's simply easier to list them separately, as Statcounter and others do.)

            Ah, on immutable distros: depending on how things roll with them, might as well argue that they're a third, different thing. I'm not knowledgeable enough on them to say anything.

55 comments