True but I hear less and less love for Ubuntu from other Linux users. Mint seems on the rise though, a lot of people hate snap and mint removing it from Ubuntu fixes that problem.
I don't like ubuntu (lots of sneaky phone home stuff, forced updates, etc) but they do provide a sort of frozen os with a stable ABI.
I honestly don't know how game developers provide binaries for linux (especially with sound and graphics). It seems to be more source compatible than binary compatible.
I suspect with ubuntu, they can just target ubuntu 14.04 or 16.04 and it will probably work with later ubuntu versions.
Can you ship a game For arch? it seems like periodically there would be a firedrill and the game would have to recompiled and an update pushed out. (or is the deal with arch that the OS adapts to the games?)
This kind of worries me, Ubuntu is by far my favorite distro, and one of the only reasons I don't just go back to windows after 10+ years of Linux, largely specifically because of Snap.
Although, I suppose if it went down some Flatpak based distro might take it's place eventually.
You actually like snap? Wow I've never heard a Linux user say that :)
I never was a big fan of Ubuntu myself also because I hate Gnome (really the whole philosophy is completely counter to my own) and Kubuntu never really worked for me. For example it didn't allow ZFS on root when I tried it last (20.04 I think). The snap thing was an extra detractor for me but I was already off Ubuntu at that point.
In the end I ended up with FreeBSD and KDE which made me a lot happier.
I've been a pretty big Snap fan ever since hearing they fixed the forced updates issue. I still wish they had an open backend, but I think traditional package management just is not a good fit for complex third party apps that need to be up to date without breaking.
It's kind of an unreasonable expectation, but Snaps(And presumably Flatpak and NixOS, if they were as popular) meet it pretty well anyway.
I'd prefer GNOME stick a little closer to the traditional desktop metaphor, but they've gotten a lot closer and more mouse friendly, and burn-my-windows is nice.
I probably wouldn't have designed a DE like that... but it seems to be "The standard Linux DE" and it seems to be extremely reliable, although that might just be because it's the first Wayland DE I've ever used.
I did always like KDE, but the older versions always seemed very slightly unpolished with something just a few pixels off from where you might think it should be, and things like that, and Ubuntus never really did seem to fully work quite right on customized DEs, without lots of work.
nagara (?), popos (?), manjaro, flatpak, and endeavor all have bigger market share than debian? Can this be correct?
Separately. Since linux distros have a hierarchical structure (A based off B based off C) these market share graphs could be smarter. Namely it'd be more useful to have "debian and things downstream to debian" represented rather than just debian itself (maybe in addition to picture we're given, if not instead).
Well they list Flatpak as a distro and I have a feeling they're including SteamOS running on Steam Decks in the stats which is based on Arch now. I'm not sure how much we should read into these stats.
If instead of desktop popularity, there were a need to run servers (non-RPi):
- Reliability, long life, low security patch, delay and maximum compatibility for enterprisey things with a faint hope of rip-replacing with RHEL for $upport (BYO RPMs of new shit): CentOS 9 Stream*
- (Same as above but it doesn't have to be Linux): FreeBSD or SmartOS
- Replace VMware: Oxide (hw+sw) or Triton Datacenter/SmartOS
- Cheap, reliable, maximum package choices: Debian
- Beginner Linux skills requiring guides from the interwebs (reliability and security updates optional): Ubuntu LTS
- Live in an empty room without furniture: Alpine
- Maximum tinkerability in a test environment without user support: Arch or NixOS
* Alma and Rocky while having a longer lifecycle promise, remain vulnerable for much longer by shipping security fixes slowly and have EL compatibility gotchas. Their engineering is also slower and less resourced.
Default to standardizing on Cent without a very good exceptional reason(tm), don't do a hodgepodge, and do configuration management.
Its not about desktop popularity or servers, its about gaming distros, or rather the packaging systems being used that the ProtonDB data seems to choke on.
So we’re measuring flatpak separately how.. if I’ve installed Steam via Flatpak instead of Pacman? So I fall into Flatpak, and not Manjaro? Am I missing something? That seems flawed.
In this case, the ProtonDB data analysis yields results that might not be specifically about the underlying distro per se and more about the packaging system used, as others here have pointed out.
