• 0 Posts
  • 117 Comments
Joined 3 years ago
cake
Cake day: June 6th, 2023

help-circle
  • tl;dr (understandable, to be honest): on a technical level, modern GNOME prioritizes polish at the expense of flexibility, and COSMIC is focused on customizability. Bad communication aside, they have fundamentally different goals and audiences.

    Acknowledging that this is a 4-year-old article, I think it’s important to read this as a very one-sided perspective. However, I am certainly not defending System76, as it does seem like some pretty poor behavior if the article is to be believed.

    I’m going to look past the issues over communication and behavior, as others have already addressed that in this thread. Other than that, it seems that the main issue is arguing over the role of GNOME in the software ecosystem. How I see this is that:

    • System76 is arguing for backwards compatibility and and more customizability.
    • GNOME is arguing for “bulletproof” theming of apps by restricting user choice and modularity.

    Honestly, I think this is pretty reflective of how the current state of the respective DEs.

    GNOME is the cleanest, most polished Linux desktop environment, if you use it exactly as the designers of GNOME envision. If you want any options outside the extremely limited set GNOME provides by default, you need to rely on extensions, which are less stable and less polished, and may or may not be updated to new DE versions.

    COSMIC is a clean-sheet implementation designed around modularity. It’s really the main thing they talk about. It has the advantage of being Wayland-only, and (supposedly) pretty much every element of the DE is modular, and there is a pretty substantial amount of customization available even in the fairly barebones 1.0 implementation.

    In terms of COSMIC “just being GNOME with extra color options”, I disagree. I really like the UI design concept of GNOME, and ten versions ago I used it all the time. However, over the last few versions it’s become very locked-down into only supporting one narrow way of using the desktop, and I need features outside that (e.g. system tray, options for window tiling, etc.). Even with ten extensions modifying the behavior – which causes stability issues when I get a new GNOME version – I still find things which bother me and are only fixable with manual dconf editing, which means I just can’t daily-drive GNOME.

    I think that’s who COSMIC is really for: someone who wants less windows-y, more intentional UI design than KDE, but with good customizability. It sucks if the creators of a pretty neat new DE were not effective participants in their previous DE, so I really hope they don’t make the same mistake with COSMIC, and manage it properly as an open source project.





  • I would also add that the more you modify the system (PPAs, packages not installed via the package manager, nonstandard partition layouts) decreases the stability of your system and makes it harder to get back to your current system state if something goes wrong. I like to think about it like balancing a tower of blocks as a kid. Mint is the first block, and is very stable, but each additional block makes the system less and less stable. Mint itself is really stable, but if you do weird stuff the Mint devs can’t do anything about it, which puts you in a bad position until you really know what you’re doing.

    The Snap store is intentionally left out by Mint, because they don’t like how Ubuntu manages it. This means that even though the Ubuntu version Mint is based on supports Snap, there’s no guarantee that snaps will work with the same stability which .deb/apt and flatpak packages will, because it hasn’t been tested in Mint. I would advise against using it.






  • I had no idea! It seems like it was a really unruly project to manage, but it’s a shame to lose the centralization of having one app that can configure anything. I don’t see any problem in having package management split off into Myrlyn, but it sounds like Cockpit is much more limited in scope, which is a shame, since handling the edge cases gracefully was what made YaST so useful.

    Here’s a source for others who didn’t realize.








  • Zettlr! Its designed around writing manuscripts in markdown+latex, then exporting to pure LaTeX, PDF, or any other Pandoc-supported format via a builtin Pandoc GUI. The only thing that doesn’t work particularly well is the table editor, but they’re working on it.

    It is electron based, but almost all graphical editors for markdown + inline latex are (obsidian, etc.) because MathJax & KaTeX are the most mature method to render LaTeX inside other document formats.

    Obsidian is also good, but it’s not FOSS and their built-in export isn’t great.