• Captain Aggravated@sh.itjust.works
    link
    fedilink
    arrow-up
    1
    ·
    9 months ago

    If I recall the “Linus killed Pop!_OS in minutes just trying to install Steam” fiasco, the forensics shook out something like this:

    1. Pop!_OS’s onboarding experience doesn’t (or didn’t at the time) walk users through a software update. At least at the time, I haven’t used Pop!_OS recently so this may have changed, but the way you would use their GUI to run the equivalent of an apt-get update was to open the Pop!_Shop to the Installed tab and…wait. So the apt cache (the local copy of the catalog of packages available in the repository) is whatever it was when the install media was created.

    2. It just so happened that the exact version of the steam.deb package that apt cache pointed to had an error in its dependencies–it claimed that it was incompatible with Pop!_OS’ desktop environment, and thus to install Steam, it would have to remove the GUI and all its dependencies right on down to Xorg. This issue was discovered and a patched version was pushed to the repository, but because of the way repositories work, you can still request an older version of software.

    3. Linus picked Steam and clicked the install button in the Pop!_Shop. It attempted the install, saw the dependency error, and bombed out, kicking up an error message “Failed to install Steam” with a bit more text.

    4. Linus Sebastian, then CEO of a technology media company, comprehensively failed to google the words in the error it gave him and find several independent forum posts, Reddit threads, and Steam community discussions saying “Run apt-get update and try it again.” Instead, he got up on his high horse about Linux GUI’s not working, started fussing about how you have to do everything in the terminal, and he instead looked up how to use the terminal to install Steam.

    5. I don’t think he recorded his screen thoroughly enough to be sure, but either the page he found was strange or he skimmed a bit too quickly. Almost all of the time, web pages containing instructions for how to install software in a Debian-based Linux system (with the apt package manager) will instruct you to run the command sudo apt-get update and probably sudo apt-get upgrade first, then probably a sudo apt-get install [packagename] None of that happened, he just found the install command and ran it.

    6. The terminal spat out it’s usual litany of “doing stuff…” before spitting out a long list of things it was going to uninstall, followed by a warning in bold allcaps to the effect "WARNING! This operation is likely to permanently damage your operating system. You should not do this unless you know exactly what you are doing. To continue, type “Yes, do as I say.” Most of the time, a Linux system requires a y or n, and might even default to y if you just hit enter. Sometimes, in order to wake people up and kind of ask “are you sure?” it will reject a simple y and tell you to type out the word yes. Asking you to type out a complete sentence is a severe warning.

    7. Linus typed “Yes, do as I say.”

    8. APT uninstalled the entire GUI and dropped into a Bash shell.

    In aviation, we talk about the accident chain. Few aviation accidents can be traced to a single brief action; instead a series of adverse events and mistakes lead up to an accident, and correcting any of them will avert disaster. Well, I count this accident chain as 8 links long. Contributing factors range from Linus’s bad attitudes and poor troubleshooting skills to the Steam package’s flawed dependencies to Pop!_OS’ flawed package manager which doesn’t refresh the apt cache on launch. The result was a crash and burn on international television.

    • Aasikki@sopuli.xyz
      link
      fedilink
      arrow-up
      0
      ·
      9 months ago

      The whole point imo was supposed to be to test linux from a point of view of a regular user, and that while surely a bit extreme, isn’t too far from what might happen when a newbie stumbles on an issue (which may or may not happen depending on luck).

      Then again he did test it on uhh, quite interesting hardware that’s almost guaranteed to have issues. Maybe it would have been more fair for him to switch to a more conventional desktop for the duration of the experiment.

      I didn’t really see that affecting the market share of Linux much either way. Luke who has a normal desktop also had his fair share of more minor issues and so did I when I ran fedora for a bit over half a year on my desktop last year ago, as a similar experiment. What has and will affect the market share, positively too, is the steam deck, which Linus is also a big fan of.

      I still use linux a lot in my daily life even if not on my desktop, my home server runs on TrueNAS scale and I have a couple projects running on raspberry pi’s (more probably coming). But for desktop use, I’ll let it cook for a couple more years before giving it another try. Running Linux on my desktop is absolutely my dream and end goal, but unfortunately it still has too many issues specifically for my (actually very broad) use case of video editing, Photo editing, 3D modeling, graphics design, gaming and more. If it was only one of those uses, I’d be able to fairly easily come by with solutions, but when it’s all of them combined, it adds up and becomes a real chore, unfortunately.

      • eskimofry@lemmy.world
        link
        fedilink
        arrow-up
        1
        ·
        9 months ago

        The whole point imo was supposed to be to test linux from a point of view of a regular user, and that while surely a bit extreme, isn’t too far from what might happen when a newbie stumbles on an issue (which may or may not happen depending on luck).

        A regular linux user is expected to be familiar with the OS

        • Aasikki@sopuli.xyz
          link
          fedilink
          arrow-up
          1
          ·
          9 months ago

          I meant as a regular computer user who’s new to Linux. Like most of them would be.