1. 12

  2. 10

    The saddest thing about this post is that it takes as axiomatic, without any thought, the idea that the core will be Linux. Today, GNOME runs well on *BSD and I think it even runs on Hurd these days but there’s a gradual trend towards Linux-only technologies. For example:

    For example, Flatpak, love it or hate it, is an easy way to install any app on any platform

    Sure, as long as your platform is either built on a Linux kernel or provides a VM that can run a Linux kernel. Otherwise… not so much.

    Linux, *BSD, Solaris, and so on are all based on ‘70s minicomputer abstractions that have been gradually stretched to breaking point for modern hardware. If you give up even supporting a handful of kernels with almost identical abstractions then you have absolutely no chance of running on anything that’s designed for modern hardware. Once the Fuchsia + Flutter stack is a bit more mature, I can see it completely eating GNOME’s lunch and GNOME at that point being completely impossible to run on Fuchsia in anything other than a Linux ABI compat layer, where it’s segregated away from any platform-integration services and completely sidelined.

    When I started writing Free Software, portability was a badge of code quality in the community. Now it’s seen as too much effort, a mindset led by companies like Red Hat and Google that don’t want developers to put effort into platforms that they don’t control.

    1. 2

      Totally agree, but this is one of those ‘inherent to human nature’ situations. Most people who run FLOSS operating systems run Linux. I’d wager that the author has VERY little experience with folks running anything but on the desktop.

      Therefore, EVERYONE runs Linux, right?

      1. 2

        You are absolutely right, and I would like to add:

        • Portability gives users more choice. If your program is their critical one, and it is portable to all platforms, they can change platforms with impunity. So the author’s very thesis argues against Linux-centrism.
        • You alluded to this, but the corollary is that the more software is portable, the more freedom users have to get out of user-hostile platforms. If all programs were portable, Microsoft, Google, and Apple wouldn’t be able to have walled gardens because if they started bring user-hostile, users would just move.

        Portability needs to make a comeback as a lauded attribute of software. I could write a blog post about this…

      2. 4

        This post frustrates me along a number of different axes, so I will attempt to keep this focused and brief.

        The author asserts that Linux is not a platform, but that Gnome aspires to be in and of itself a platform. That is, it wants to be more than just a set of libraries and SDKs, but also to embody best practice, interface guidelines, and one would hope interoperability guidelines to allow apps to more productively work with each other, the operating system, and the user.

        This is a laudable and lofty goal, but please forgive me if I am unwilling to believe. The experience I have had with Gnome through the years is that while other desktop environments like KDE quietly chug along, maturing over time, Gnome keeps blowing up the world and starting from scratch every few years. This results in critical polish that is absolutely essential for some people to use their computers is absent.

        Case in point, this bug while only reported 7 months ago in this particular incarnation, has meant that ever since ubuntu 19.10 shipped I am unable to use Gnome as a visually disabled person because the Zoom accessibility feature, table steaks on every popular platform, leaves square mouse trails as you use it when not currently zoomed in.

        You can’t build a house if the foundation is rotten, and I would argue that rather than trying to aspire to build a platform, Gnome should focus on structural soundness and the quality of its foundation.