• 0 Posts
  • 732 Comments
Joined 2 years ago
cake
Cake day: July 16th, 2023

help-circle

  • I agree with you completely. I am sure you deal with these minor issues quickly and barely notice them half the time.

    But users of other distros would find it intolerable to have to deal with these small tweaks on any given day. “My computer is a tool” they will say and “it just needs to work”.

    Fair enough. But then they turn around and fight bugs and limitations that were solved for Arch users months or even years ago.

    And they fight to install software not in the repos, often making their overall system less reliable in the process.

    I prefer the stability of Arch over the stability of Debian thank you.


  • The problem is that “stable” means two different things in Linux.

    It can mean “reliable” as in it does not crash. I think that is what most of us think of.

    However, It more often mean “static” or “unchanging”.

    Take Debian Stable. It is “stable” because the software versions rarely change outside of security updates. This does not mean it does not crash. It does not mean it does not have bugs. It means you can depend on it to behave tomorrow like it does today. Design problem not the software installed? They are not getting fixed. As an example, you will see that the people saying Wayland does not work are almost always Debian users because they are using software from 2 - 3 years ago. Debian 13 has improved things but the NVIDIA drivers are from 2 years ago even now. And if KDE has fixed a lot of bugs, that does not mean Debian gets those updates.

    Arch on the other hand updates its packages constantly to the latest to very recent versions. The behaviour of your Arch system changes all the time as new versions of software are installed. You may like this or you may not but this is “unstable” using Debian’s definition.

    From the point of view of robustness, Arch users often have a better experience than Debian users. Things more often “just work” due either to new features or because issues have been resolved in recent versions. Rapidly developing software, let’s take Wayland or NVIDIA again, will often work dramatically better on Arch. However, every update has the potential to break something. And so, on Arch, you are certainly more likely to encounter breakage. Often these problems are very short-lived with fixes appearing quickly. This means that, even if something did break, many Arch users will not even know.

    Anyway, this is my take Arch vs Debian:

    • Arch is more “robust” (fewer problems on a typical day)
    • Arch is very reliable but less reliable than Debian (updates rarely break but they can)
    • Arch behaviour changes much more often (more features sooner but also more learning required and occasionally features lost or “get worse”)

    So, it all depends on what we mean by stable


  • Mint has two kernels: a “stable” one and a “hardware enhanced” one (HWE). The HWE kernel is newer to improve support for newer hardware.

    Many distros allow you to pick from multiple kernels.

    Yes, all Linux kernels come from kernel.org

    That said, kernel.org maintains not only a latest but also multiple “stable” kernels that maintained versions of previous kernels. There are usually about a half-dozen kernel versions to choose from.

    One you have code from kernel.org, you can change the configuration to get kernels with slightly different capabilities and strengths.

    Finally, you can patch the code you get from kernel.org to add or remove whatever you want. For example, you may add in filesystem support or drivers missing in the mainline kernel.

    So, in the end, any given Linux distro may have a Linux kernel slightly different from what other distros use. You can probably run any Linux distro on the kernel from any other Linux distro though. In fact, this is what you are doing when you run something like Docker or Kubernetes.


  • Governments do not have to be involved in projects to pass laws that impact them.

    I would argue greater EU participation in FOSS would improve the situation. One, the number of people in the government that understand how FOSS works may increase and frankly ignorance is often the problem. Second, if lawmakers themselves or the things they care about rely on FOSS, they will be much less likely to kick the legs out from under it.

    From a code perspective, the risk is low. If it is just that they add back doors (not because it is the law), we simply create versions without those back doors and use that instead.

    I do not think that developers have any greater insight into social or legal issues than you do.
















  • Microsoft will for sure benefit here.

    Many users, especially businesses, will simply upgrade.

    Some will pay for the ESU.

    Some will sign up for cloud backups.

    All these benefit Microsoft.

    Some fairly small number will work around Microsoft’s plan by upgrading Windows 11 where they are not supposed to or finding a way to get the updates for free.

    Sure, probably the biggest fraction of users will probably do nothing. But they were already doing nothing for Microsoft so nothing changes in this case. Of anything, the load in Microsoft servers goes down a bit.

    So ya, Microsoft has little incentive not to charge ahead.