Do you use Linux?

[DrCR]

I recently got to geek out a bit with some engineers about Linux and why Ubuntu will basically never be used (industrial environment) by them vs Red Hat.

I haven’t touched an rpm distro since trying Fedora Core 2 forever ago. And I skip on it almost immediately since it seemed all wrong vs the Slackware I cut my teeth with. It basically has been Debian or Slackware based distros ever sense for me. (I tried Arch when it was a new distro, but it’s not for me.)

I have tinkering with Fedora (maybe Fedora Spotlight, whatever twist that is) and Rocky on my to do list now due to that conversation, but I’m curious— am I the only one here who found/finds Fedora and co weird?

 

[SchumannFrequency]

Check Steam's hardware survey, which is more relevant considering ProtonDB's stats are most likely a fraction of what Steam has, given that the former's users are subset of the latter.

You're talking about sample size here.
It's clear that you think 'The steam hardware survey' has the larger sample size. The problem with your idea is that there isn't any hard evidence for this. If you ever followed the GamingOnLinux website, you would know that this survey is frequently discussed here and its reliability is questioned by many/most users. Mainly because most people seem to have never received this survey.

Steam Hardware Survey For January 2015

Here is the Steam Hardware survey for January 2015, starting the year with not many changes as expected.

www.gamingonlinux.com

As always, remember this is a survey, so it won't ask every single one of you to do it. It would only be truly accurate if it did it behind the scenes, but that's not what a survey is for this is just to get a general idea.

The steam hardware survey bias against linux
https://www.reddit.com/r/linux_gaming/comments/7h48rw I am using linux since more than 2 years, with steam open nearly every day. I didn't get the survey once. Now I set up a windows VM with GPU passthrough and all the nice stuff, and after a few weeks of opening steam there, it asks me for a hardware survey on windows. What the fuck steam?

Now consider: A: A website for enthusiasts and users who are more technically oriented, and B: a service that practically every gamer subscribes to.
A relies on conscious participation and self reporting, while B is -partially- automated.
A samples forum users, B samples gamers directly, regardless of whether they even know what the hell a forum is.
So, which, from your experience as a trained statistician, would provide a, shall we say, "less" biased sample, assuming -for the sake of argument- ceteris paribus?

You forget that the Steam Hardware Survey is too selective to give an overall picture. Not every gamer uses Steam, and this is probably more the case with Linux gamers than with other types of gamers. You have the Linux gamers who always stay with GOG. Some only use Lutris without ever using steam. Some use wine + dxvk without ever using steam.

The GamingOnLinux users is a more 'general' sample that does not have this particular bias. They don't have to use steam per se..

The three different sources (GOL, steam hardware survey, ProtonDB) all confirm that Ubuntu is no longer the most popular Linux distro in their statistics.

"SteamOS Holo" 64 bit 24.98%
Ubuntu 22.04.1 LTS 64 bit 12.27%
"Arch Linux" 64 bit 9.12%
"Manjaro Linux" 64 bit 6.92%
Freedesktop.org SDK 22.08 (Flatpak runtime) 64 bit 5.71%
Other 40.99%

SteamOS Holo and Manjaro are derivatives of Arch Linux. If we want to calculate the percentage of Arch-based distros, we see the following:
Arch-based distros: 41.02%
Ubuntu 22.04.1 LTS: 12.27%

You also have EndeavourOS and a number of other distro's based on Arch. You have Fedora and Nobara Project that are gaining popularity.Those are not listed. And of course you also have other Ubuntu versions and derivatives of Ubuntu that are not specifically mentioned. You wouldn't be a useful statistician. The survey you always defend as the 'most valuable' is in reality a very incomplete statistic that says almost nothing at all.

I have no problem with you using Ubuntu or anything else. I think that's the advantage of Linux that everyone can choose what they prefer in many more areas. But I started on Linux sometime in 2008 and then used Ubuntu for years, and for the last few years I've still tried to use Ubuntu on my netbook and laptops, but I've now gotten to the point where it's no longer on any system .

I would recommend people who have medium or high quality requirements to leave Ubuntu for what it is and give the following systems a chance:
Nobara Project, Void Linux, Mint, EndeavourOS, MX Linux, NetBSD, Devuan, FreeBSD and Clear Linux

 

[The King]

Technically almost "everyone" uses Linux everyday, they just don't know that they do!

Example almost all routers run Linux. Numerous other devices that people encounter in their daily lives mostly run on Linux.

 

[Shihab]

The steam hardware survey bias against linux
The steam hardware survey bias against linux from linux_gaming I am using linux since more than 2 years, with steam open nearly every day. I didn't get the survey once. Now I set up a windows VM with GPU passthrough and all the nice stuff, and after a few weeks of opening steam there, it asks me for a hardware survey on windows. What the fuck steam?​

I've personally was asked to participate in Steam's hardware survey at least once in every distro I've installed it on.
1, 2, 5 or even a 100 complaining they haven't received requests out of a million is hardly something to focus on. Good samples are randomly selected. It would thus mean, by virtue of being randomly picked, you don't get a 100% chance of being included, no?
​

You forget that the Steam Hardware Survey is too selective to give an overall picture. Not every gamer uses Steam, and this is probably more the case with Linux gamers than with other types of gamers. You have the Linux gamers who always stay with GOG. Some only use Lutris without ever using steam. Some use wine + dxvk without ever using steam.​

True, but, for the lack of any reliable insight on the actual segmentation of the gaming market, we can only infer distribution, and hence significance of other stores, using other (platform-agnostic) measures. Let's try revenue. For FY2020, gog's revenue was at ~US$71mill (page 62), Steam had US$7.3 billions, that's ~100 times more revenue. I agree it's not an exact measure for the size of the user base, but it is a good indicator. And then you have -when you consider that only Steam support linux- native support, reduced complexity, low bar of entry, etc, etc.

I do not deny that there are probably many who game outside the steam ecosystem, but there are two factors to consider here: 1- It's not mutually exclusive. ProtonDB itself is a good evidence of that (database includes gaming on Wine/Bottles/Lutris, yet all its users are also Steam users). 2- Steam can function as a launcher for non-steam games.

I cannot find any stats for Lutris, and would appreciate reliable ones for it. As for manually setting up environment for gaming, I don't think we need to point out why those users would be a minority...
​

The three different sources (GOL, steam hardware survey, ProtonDB) all confirm that Ubuntu is no longer the most popular Linux distro in their statistics.​

I did state that Steam's offerings are putting the screws on Ubuntu. I avoided focusing on it because it's -probably- entirely driven by the Deck's adoption, which would make it moot to compare to desktop distros installed by conscious user choice (i.e. all the other distros we're talking about). ProtonDB does the same by treating the Deck as it's own niche with its own grouping (which is annoying, when you're trying to check how a game performs on the desktop. Have to scroll past a dozen of posts to get there! -_-). The GamingOnLinux dataset counts SteamOS separately.

I think it's unwise to start moving goal posts from talking about specific distros to "it's based on x" and then grouping several different products against one. This would, for starters, make (even more) moot your previous comparisons between Mint and Ubuntu. Heck, I could just say this entire discussion is pointless, and the most superior distro overall is Debian!
​

You wouldn't be a useful statistician.​

I know.​

 

[claes]

Heck, I could just say this entire discussion is pointless, and the most superior distro overall is Debian!

This seems like the correct answer? I honestly don’t know what Schumann’s argument is anymore but admire your patience and thoughtfulness in entertaining this… discussion

 

[lexluthermiester]

This seems like the correct answer? I honestly don’t know what Schumann’s argument is anymore but admire your patience and thoughtfulness in entertaining this… discussion

I'm having no issues following along. I think it's been interesting and insightful from both parties. Schumann has stated a few things I didn't know about and went to read up on. But I've also enjoyed reading Shihab's perspectives as well. I disagree that this discussion was pointless.

 

[claes]

I understand the words, but Shihab is right about the shifting of goal posts. It is no longer an argument about the quality of Ubuntu for desktop users but a one-sided “discussion” of which Linux distro is best for gaming, which was not the premise. And, well, the statistics are really bad, but I guess that depends on what your goal post is.

Glad you’re having fun following along.

 

[mplayerMuPDF]

You're talking about sample size here.
It's clear that you think 'The steam hardware survey' has the larger sample size. The problem with your idea is that there isn't any hard evidence for this. If you ever followed the GamingOnLinux website, you would know that this survey is frequently discussed here and its reliability is questioned by many/most users. Mainly because most people seem to have never received this survey.

Steam Hardware Survey For January 2015

Here is the Steam Hardware survey for January 2015, starting the year with not many changes as expected.

www.gamingonlinux.com

As always, remember this is a survey, so it won't ask every single one of you to do it. It would only be truly accurate if it did it behind the scenes, but that's not what a survey is for this is just to get a general idea.

The steam hardware survey bias against linux
https://www.reddit.com/r/linux_gaming/comments/7h48rw I am using linux since more than 2 years, with steam open nearly every day. I didn't get the survey once. Now I set up a windows VM with GPU passthrough and all the nice stuff, and after a few weeks of opening steam there, it asks me for a hardware survey on windows. What the fuck steam?



You forget that the Steam Hardware Survey is too selective to give an overall picture. Not every gamer uses Steam, and this is probably more the case with Linux gamers than with other types of gamers. You have the Linux gamers who always stay with GOG. Some only use Lutris without ever using steam. Some use wine + dxvk without ever using steam.

The GamingOnLinux users is a more 'general' sample that does not have this particular bias. They don't have to use steam per se..

The three different sources (GOL, steam hardware survey, ProtonDB) all confirm that Ubuntu is no longer the most popular Linux distro in their statistics.

"SteamOS Holo" 64 bit 24.98%
Ubuntu 22.04.1 LTS 64 bit 12.27%
"Arch Linux" 64 bit 9.12%
"Manjaro Linux" 64 bit 6.92%
Freedesktop.org SDK 22.08 (Flatpak runtime) 64 bit 5.71%
Other 40.99%

SteamOS Holo and Manjaro are derivatives of Arch Linux. If we want to calculate the percentage of Arch-based distros, we see the following:
Arch-based distros: 41.02%
Ubuntu 22.04.1 LTS: 12.27%

You also have EndeavourOS and a number of other distro's based on Arch. You have Fedora and Nobara Project that are gaining popularity.Those are not listed. And of course you also have other Ubuntu versions and derivatives of Ubuntu that are not specifically mentioned. You wouldn't be a useful statistician. The survey you always defend as the 'most valuable' is in reality a very incomplete statistic that says almost nothing at all.

I have no problem with you using Ubuntu or anything else. I think that's the advantage of Linux that everyone can choose what they prefer in many more areas. But I started on Linux sometime in 2008 and then used Ubuntu for years, and for the last few years I've still tried to use Ubuntu on my netbook and laptops, but I've now gotten to the point where it's no longer on any system .

I would recommend people who have medium or high quality requirements to leave Ubuntu for what it is and give the following systems a chance:
Nobara Project, Void Linux, Mint, EndeavourOS, MX Linux, NetBSD, Devuan, FreeBSD and Clear Linux

My experience with NetBSD on AMD64 (multiple versions and computers) was absolutely horrendous. I don't remember the specifics but it mostly had to do with the installer being fatally buggy. Really wanted to like it and never had an issue with on a Raspberry Pi B+ (armv6hf) but it was not to be. FreeBSD has serious quality issues too. If you want a quality *BSD then OpenBSD is the best option but using it as a desktop OS (which I personally did a couple versions ago for some months) comes with its own challenges.

 

[SchumannFrequency]

My experience with NetBSD on AMD64 (multiple versions and computers) was absolutely horrendous. I don't remember the specifics but it mostly had to do with the installer being fatally buggy. Really wanted to like it and never had an issue with on a Raspberry Pi B+ (armv6hf) but it was not to be. FreeBSD has serious quality issues too. If you want a quality *BSD then OpenBSD is the best option but using it as a desktop OS (which I personally did a couple versions ago for some months) comes with its own challenges.

NetBSD is relatively good. But this is not something that applies in every situation and for every user. Two years ago I had several European wi-fi USB sticks that didn't work on FreeBSD. In itself this is not a super big problem because I also had a (cheap) extender that I could connect via an ethernet cable.

I had then tested NetBSD and all these wi-fi USB sticks were working, and I've only used NetBSD for a few days but had no problems with anything. NetBSD's wi-fi USB support is sometimes better than FreeBSD, although I don't know if it still applies as it's been worked on for the last year.

Another NetBSD advantage:

https://twitter.com/i/web/status/1355017418238816257

on a lightly loaded NetBSD system, compute bound code is 10%+ faster than on a Linux box with same CPU. One big difference, the number of involuntary context switches is much higher on Linux - maybe due the hundreds of kernel processes in Linux.

In fact, I consistently show #netbsd runs applications faster than Linux - these are not systematic tests, but Linux has a huge overhead now from the vast number of system processes running all the time

It's always said that installing NetBSD is a good learning experience and things like that, and that the installation can be difficult, but I don't know. I didn't find installing NetBSD especially more difficult than other systems. It hadn't occurred to me. The Register recently reviewed it: https://www.theregister.com/2022/08/10/netbsd_93/

The above article is completely subjective. He says things that he actually contradicts in his article later. He could have just installed XFCE directly and then most of his article no longer applies..

Another benefit of NetBSD is security so it is arguably the most underrated system in the universe.

OpenBSD excels in hardware support and security, but has the major flaw that it only has half the performance of FreeBSD. FreeBSD is already a very secure system and can be made as secure as OpenBSD with some extra work.

In terms of quality, I don't think there is anything better than FreeBSD. So I don't know what problems you experienced.

 

[DrCR]

was cleaning out some of my notes and came across this link. Not sure where I came across it (it may have been this thread), but posting here to share since I found it an interesting read.

https://ludocode.com/blog/flatpak-is-not-the-future

 

[Easy Rhino]

was cleaning out some of my notes and came across this link. Not sure where I came across it (it may have been this thread), but posting here to share since I found it an interesting read.

https://ludocode.com/blog/flatpak-is-not-the-future

Yes, there have been a lot of hot takes around flatpak over the years. I think the concept is the future but the implementation needs work.

 

[SchumannFrequency]

Yes, there have been a lot of hot takes around flatpak over the years. I think the concept is the future but the implementation needs work.

In terms of boot times, performance after boot, and disk space usage, AppImage outperforms Snap and Flatpak.

Somehow, it's a bit odd that Canonical and Red Hat don't just have a chat about using AppImage together. Thus finally solving the old compatibility problem with something that is most acceptable. You also have GNU Guix and Nix which may also be a better concept than Snap and Flatpak.

So that's a strange phenomenon to watch. That two companies you would expect to have smart engineers fill their days pushing inferior tech and making it the norm. A bit like MS.

 

[DrCR]

smart engineers fill their days pushing inferior tech and making it the norm. A bit like MS.

Well, Red Hat will Red Hat. (We could talk init systems…) And Canonical is too small of a player comparatively.

 

[vectoravtech]

I use linux mint mate in a dualboot with windows 11 for gaming.

 

[Goodman]

I choose as my main operating system even if I'm using "dual boot" on separate SSD (just choose in Bios which one I want to boot from)
But I rarely boot into windows maybe 2 or 3 times a year lol! (Win7 Ultimate)

True that I rarely game anymore so windows is not important to me but my Dos games works super fast in Linux joking! , quite frankly except for a very small number of games pretty much all my games works good in Linux using Wine , Playonlinux , Steam/Proton , Lutris , etc...
If you got a good "recent" PC 5 years or less you could probably play pretty much all the (recent) games out there in Linux specially if you got an AMD GPU with Nvidia you may be a bit faster but you will have to install their proprietary drivers and you'll get more games not working with Nvidia than you would with AMD GPU because of AMD open source drivers

Anyhow I tried Linux in mid 90's but was to hard (Red Hat) then I give up until I tried Ubuntu in around 2006 or 2008 much easier but I didn't like the Unity thing moved on to Mint liked it a lot use it for a while then tried some more distro's like Arch , Debian , or Ubuntu based , you name it I probably tried it before also different DE , XFCE , Mate (really don't like it) Cinamon , Budgie and now using only KDE that is best for me

What got me into KDE was Feren OS easy to use and learn and quite frankly probably best KDE out there but it is often to late in the kernels and some packages so now i'm using only Kubuntu on my main PC there is also KDE Neon of course but to "lite" for my taste

KDE is so nice and so much customization that you can do with it that once you are used it , using anything else seams like you're missing on a lots of things that you can do with a distro and you fell like going backwards with any other DE but that just my opinion

 

[Icon Charlie]

The old man with boat sinker jumped into the Deep Penguin Pool of the Linux... I'm too damn old and too damned tired for doing anymore code money work.

The distro that I have chosen is Zorin 16.2 and because of my IcyDock setup, All I have to do is swap SSD drives with different os's and its on to play time with Gonzo the Guerilla.

Oh yea... I'll be in here asking for help... that is for sure... but I have already seen the warning signs of Micro$haft.

And it's not good. So I better learn to use the easy distro version of Linux before Fucking Gates decides to make the OS a closed garden.

And IMHO this is going to happen... Not right now for a couple of years from now. But mark my words it is going to happen.

 

[sepheronx]

I decided with a dell t5810 sitting idle, I installed Pop OS.

So far with lutris it takes absolutely forever to install a GOG game. And installed a few games with 1 still giving me trouble.

I also had to install rustdesk via terminal and it doesn't create an executable. If I want to run it, I have to do so via terminal. Anyway to fix that?

 

[Shihab]

So far with lutris it takes absolutely forever to install a GOG game.

Try Bottles. Tested with a couple of offline gog installers, finished as quickly as they typically do on Windows. Dunno about Galaxy installs.

I also had to install rustdesk via terminal and it doesn't create an executable. If I want to run it, I have to do so via terminal. Anyway to fix that?

You mean an icon you click on on the launcher? Those are just plain text files named something.desktop in a standard directory, ~/.local/share/applications or /usr/share/applications/. You can manually create one for your program following the formula of other .desktop files you find there.

Desktop entries - ArchWiki

wiki.archlinux.org

set exec= to whatever terminal command you use to launch the application.

 

[SomeOne99h]

I found out upon many frustration experience that Flatpack version of Bottles is annoying (at least for me). I prefer the compiled package unless I feel very paranoid about security. With its sandbox-by-design going on, it gave me issues. Like I would have an RPG maker runtime installed, but when I try to run a game that needs it, it complains of missing run-time lol. Upon searching through their forum, it came down to the Flatpack package needing further adjustments, and even with that, it might still cause issues.

However, last version I used was the 8th or the 6th month version, I saw them talking about fixing those issues beforehand so I don't know.

 

[Easy Rhino]

I found out upon many frustration experience that Flatpack version of Bottles is annoying (at least for me). I prefer the compiled package unless I feel very paranoid about security. With its sandbox-by-design going on, it gave me issues. Like I would have an RPG maker runtime installed, but when I try to run a game that needs it, it complains of missing run-time lol. Upon searching through their forum, it came down to the Flatpack package needing further adjustments, and even with that, it might still cause issues.

However, last version I used was the 8th or the 6th month version, I saw talking about fixing those issues beforehand so I don't know.

Have you tried Flatseal for typical flatpacks? Also, yea I almost always install the DNF unless I am paranoid about security.

 

[pavle]

I sometimes use Linux Lite in on my computer, it's very snappy and runs well on all kinds of hardware (installed it also on an old laptop and works great).

 

[Shihab]

Flatseal is a must have for managing flatpaks.
Now if they just figured out a way to spare me >600MB downloads every bloody time Nvidia adds a 0.0001 to their system driver version!

 

[sepheronx]

I honestly haven't used linux since like fedora core 2? That was a long time ago.

I have since installed Manjaro Linux which has even less software ready to use than pop OS. I'm just testing them out right now on a Dell t5810 with a rx 6600.

Manjaro is kinda nice in Gnome desktop environment. I just got bottles installed via terminal. Now trying to figure out how to get heroic game launcher to install.

I guess that's still the reason why Linux really is not used often. Need things I stalled and done? Look up videos and tutorials just to install a program. Can't simply download either, gotta use terminal to get it to work.

I'll admit though, I'm liking terminal and enjoying my time learning it. It's just so many will not and thus I can understand people's reluctance. Been so long I forgot everything.

What I do love is how the amd display drivers are just automatic and doesn't require constant installs. Instead it's done via overall system updates.

Damn. Manjaro screwed itself. PC just wouldn't power back up. Wiped the drive and did Pop OS again.

My only qualm is I couldn't get rust desk to work in Pop OS. That's why I tried Manjaro. I mean I could get rust desk to work but I had to rebuild it every time if I wanted to.

 

[bug]

I honestly haven't used linux since like fedora core 2? That was a long time ago.

I have since installed Manjaro Linux which has even less software ready to use than pop OS. I'm just testing them out right now on a Dell t5810 with a rx 6600.

Manjaro is kinda nice in Gnome desktop environment. I just got bottles installed via terminal. Now trying to figure out how to get heroic game launcher to install.

I guess that's still the reason why Linux really is not used often. Need things I stalled and done? Look up videos and tutorials just to install a program. Can't simply download either, gotta use terminal to get it to work.

I'll admit though, I'm liking terminal and enjoying my time learning it. It's just so many will not and thus I can understand people's reluctance. Been so long I forgot everything.

What I do love is how the amd display drivers are just automatic and doesn't require constant installs. Instead it's done via overall system updates.

Damn. Manjaro screwed itself. PC just wouldn't power back up. Wiped the drive and did Pop OS again.

My only qualm is I couldn't get rust desk to work in Pop OS. That's why I tried Manjaro. I mean I could get rust desk to work but I had to rebuild it every time if I wanted to.

There's a distinction to be made here.
Most software for Linux is available from an official or a 3rd party repository. One line to install, updated with rest of the system forever.
What you are describing is closed-source software that is not possible to host in a repository, unless the software maker decides to host their own (which some do, but that's the exception rather than the norm).

The whole package management was even copied by Windows as winget.

 

[sepheronx]

There's a distinction to be made here.
Most software for Linux is available from an official or a 3rd party repository. One line to install, updated with rest of the system forever.
What you are describing is closed-source software that is not possible to host in a repository, unless the software maker decides to host their own (which some do, but that's the exception rather than the norm).

The whole package management was even copied by Windows as winget.

Thing is rust desk is open source. Problem is that pop OS is missing dependencies that the program requires so it can't install. For life of me I can't get the nightly build to install that removes the requirements of pulseaudio.

I'm still learning terminal as it's been so long, years of alcohol abuse and age creeping up that my brain doesn't quite work as well it used to.