The Linux Thread - The Autist's OS of Choice

it's why X windows was designed to allow windowing over the network, back when dinosaurs still roamed the earth and Solaris was relevant
X being a client-server architecture was always a dumb idea that has made for unnecessary complication.

Why Wayland decided to copy that (and add in 100 other retarded ideas) is beyond me.
However, with the rise of AI, a workload that most people, on most hardware, will simply never be able to run on their own device and still do anything else useful
I don't think this is accurate, already Apple runs many smaller AI models directly on-device including iPhone to do many useful things like searching photos by keyword, predictive suggestions, text highlighting in images, and etc. Who would have thought 16GB of RAM in a phone would be "normal" even a few years ago? Continuing to 128GB and current laptop GPU performance in a few years would make ChatGPT level LLMs viable, which isnt far fetched.
that world may soon become a reality
A better analogy would be web browsers which became the modern thin-client. Unfortunately now web pages demand oodles of RAM and CPU time to run Angular/React/NextJS/other-shitty-framework so I don't know how true that is anymore.
 
Gamers just want good products, migration to the broader linux sphere from SteamOS will be minimal because if SteamOS "just works" as a gaming pc then they will have no desire or need to move, and if they feel no need they wouldn't really give the idea of linux itself much thought. Really if I was a betting man I'd say that if the valve gaming pc takes off then SteamOS will just end up in its own ecosystem with a somewhat different culture to the typical linux scene, though since any work on SteamOS would be work on Arch (thus work on linux) this would be fine too.

The benefits that have actually come out as a result of valve working with linux is Proton+Wine improvements, outside of that there is nothing special about the whole SteamOS situation, other than the fact that (big company) worked on it. As a whole it feels like this is what people want valve to do judging by public sentiment, make a distro and slap their name on it to appeal to normies and the like, but I don't see how this actually improves linux itself, and its in practice very akin to asking for a corpo bail out or hand me out, the best case scenario for any corpo working with linux is that they can provide software that the broader community can use to improve linux further.
I think it'll be because windows becomes completely unusable for normies
People will get over ai in their os as millennials grumble but if Microsoft fails to make a product people want on their PCs then Linux will gain mainstream attention
 
X being a client-server architecture was always a dumb idea that has made for unnecessary complication.

Why Wayland decided to copy that (and add in 100 other retarded ideas) is beyond me.
Being client-server wasn't the problem, it was trying to do it over the network that was the problem. Doing windowing over a socket is just how you would do it on a Unix OS. Even if you provided it to the user via some library abstraction, there would still be some kind of IPC under the hood with processes drawing their own windows and some central process managing putting them all on the screen together. It's just the sensible way to design it.
Wayland's problems aren't really in trying to do too much, they're in trying to do too little, leaving too much up to specific implementations/extensions of the protocol, resulting in lacking features and incompatibilities.
 
It should be noted Microsoft themselves has pushed hard for Office 365 over running Office natively, which only requires a working web browser, and plenty of other enterprise software companies have been pushing Cloud-based/browser native versions of their software for a while now.
The online version of Excel still doesn't have feature parity and never will because you'd need to let a sheet execute arbitrary code...and I think everyone knows that is a bad idea. This is the biggest stumbling block.

Wine doesn't do well with the newer versions of Office since they've got hooks deep in the kernel for performance reasons.

Now, if you made a Linux distro that'd make it easy to put your existing Windows install into a VM then that'd be a game changer.
 
Now, if you made a Linux distro that'd make it easy to put your existing Windows install into a VM then that'd be a game changer.
"WinBoat" doesn't really count for your existing install. But they do the 97% compatibility thing by making Windows apps easy to run in a VM. Not sure if they have GPU stuff working properly yet or not.

"WinApps" did seem similar but seems defunct.
 
This is the biggest stumbling block.
Whenever there is a need for Excel VBA, we're talking about some decades old script that is complete black magic, written by someone whose identity has been long lost in the company records, that just works and is expected to work for an eternity. You won't move businesses that rely on those to Office 365, and you sure as hell won't move them to LibreOffice.
Now, if you made a Linux distro that'd make it easy to put your existing Windows install into a VM then that'd be a game changer.
Meanwhile in the real world, not the one built in the minds of Linux cultists, this is one of the most ass-backwards ideas to have ever, given how the context of this statement is fucking Office of all things. Serves zero purpose whatsoever, besides the fact that now you run Linux at the bare metal, so that's +10 inches to your ego cock.

If all you do is browse the web and game, Linux already handles it pretty well without a VM.
If you have some software that needs a Windows environment (Adobe, Autodesk), it's a better idea to set up a clean VM for those.
If you want this to be used in those places where you need stuff like Excel VBA's in hopes it'll somehow make Linux adoption more viable, which is what I gathered from the context of the post, if you work at IT in a company like that you should be fired and blacklisted for coming up with such useless ideas that will do nothing but make everyone's work more miserable because you've overdosed on ideological bullshit that made you forgot what software actually means at it's core, making you a liability to anyone willing to hire you at any position involving something that can run Linux.
If you are a Windows user that wants to try Linux, you know how everyone did it before the advent of all the Winboat KVM fuckery? Dual boot. Just dual boot like everyone did for fucking decades and stop being a nigger trying to reinvent the wheel. Look at Wayland, that's what you get for reinventing the wheel, a heap of shit that sets back desktop Linux by decades.

If anything the ability to cleanly run an existing Windows installation within a KVM where you can then rawboot back into it would be the thing you'd like to add to a distro, so that if you need to pull something from your Windows install, you can just run it transparently within Linux and not have it shit itself once you run it outside a VM environment. But this wouldn't be some silver bullet to getting Windows people to move over to Linux (because that's always what's being looked for with these solutions, let's be brutally honest), however it will be immensely helpful to people who are in the process of your classic dual boot Linux migration.

Just, stop trying to find some magical solution that'll make Linux experience transparent to your regular Windows user. That's just trying to shove a square peg into a round hole because you have a blindfold and can't think of anything better.
 
As a whole it feels like this is what people want valve to do judging by public sentiment, make a distro and slap their name on it to appeal to normies and the like, but I don't see how this actually improves linux itself, and its in practice very akin to asking for a corpo bail out or hand me out, the best case scenario for any corpo working with linux is that they can provide software that the broader community can use to improve linux further.
This is how things happen on Linux. I by no means think Valve with revolutionize linux, or anything like that. But having them make a project like that would likely be an overall benefit to the linux community, since that's basically how things have improved through the history of linux.

A company wants to make a product, or provide a service. They choose linux for any number of reasons. Then when they run into problems they either work with, or provide code for the kernel to fix the issues they have. Or to get the functionality they need added.

Valve undertaking making a desktop gaming operating system, would likely lead to them finding things they want to add, or fix, so the product they provide won't give their users any issues. It could end up with them implementing their own wayland protocols, that end up getting accepted despite the people that are running the wayland project's constant stalling over nothing arguments. Providing bugfixes to KDE if needed, which I'm sure there are some KDE bugs that need it. Since there always seems to be. You get the idea.

Then there is the effect of having a larger userbase on linux. More people using it means, more people want to make software, and hardware around it. But it brings the negative consequences of a larger user base also.
 
Here's a blast from the past I wasn't expecting to reminisce about: I was reading the Wikipedia article for GNOME Videos (formerly Totem) and in the history section, I couldn't help but notice there were two versions: one that used GStreamer (tried and true) and another that used xine. Holy fucking shit, I have not seen the word "xine" in like... 13-14 years. I always jumped straight to GStreamer because I understood GStreamer as "the thing that pulls in all the codecs necessary for me watch pirated Fullmetal Alchemist episodes and Jimmy Eat World MP3s." I mean... it's all moot anyway because MPV exists now, and even before MPV, GStreamer, MPlayer, and VLC existed and had way more features (including NPAPI browser plugins!). "xine" exists in my mind as "that media framework that exists parallel to GStreamer but no one ever uses." Apparently, it's still active and using Sourceforge as its repository... last update was two days ago. That said, most recent comment on the repository was back in 2013.
 
Similarly, the reason Blender became such a magnificent piece of software is because everyone was sick of Autodesk and 3DS Max being a steaming pile of shit. That's when a massive development push happened when we got Blender 2.8 and all the subsequent versions where Blender is now the industry standard and it's completely free.
It isn’t, to my knowledge nobody serious actually uses blender. It might be a good pice of software but it doesn’t come with a warranty or 24 support and other things professional shops care about. That’s another barrier for FOSS software unfortunately.

Granted for Linux RHEL exists for this exact reason and what our wonderful community fucking despises them for it.
 
Now I don't know if this is because of better Kernels, Omarchy, Arch Linux or just because its the Year of the Linux Desktop™(fr this time guies) but I had to quickly print something and for the first time in my Linux life, I pressed print on my computer and just werks™... well granted the fonts were fucked, but thats never happened to me on Linux, especially when my printer is on a wifi network.
 
It isn’t, to my knowledge nobody serious actually uses blender. It might be a good pice of software but it doesn’t come with a warranty or 24 support and other things professional shops care about. That’s another barrier for FOSS software unfortunately.

Granted for Linux RHEL exists for this exact reason and what our wonderful community fucking despises them for it.
Kinda recently, there was the film Flow that won both a Golden Globe and an Academy Award, and it was entirely made in Blender. There are examples of large companies using it for part of their workflow, usually a small part like using a single tool that Blender has that other options just don’t, and it’s used as the main tool in some universities for teaching 3D modelling.

The whole “it doesn’t have a warranty so corps won’t use it” is true to an extent, but it’s also a bit of a cope in a lot of cases. Sometimes the corps use what they use because it works better than your hobby project.
 
Now I don't know if this is because of better Kernels, Omarchy, Arch Linux or just because its the Year of the Linux Desktop™(fr this time guies) but I had to quickly print something and for the first time in my Linux life, I pressed print on my computer and just werks™... well granted the fonts were fucked, but thats never happened to me on Linux, especially when my printer is on a wifi network.

Printing on Linux actually benefitted from Apple of all companies. I believe it was Apple that created CUPS (the Common Unix Printing System), released it as open-source software, and it basically became bog-standard on everything from OSX to Linux to BSD and even whatever proprietary Unices that still technically exist like AIX, HP-UX, and Solaris. USB printing was always seamless for me, even like 14-15 years ago when I first started fiddling with Linux. I never had a network printer until recently, and as far as I could tell, the process was seamless. Printed out test pages just fine, scanner recognised, all that good stuff. I should point out that this was with the aid of GUI tools in MATE, Cinnamon, GNOME Shell, and Plasma 5.

I think on more minimal setups like base Arch, Debian netinstall, Gentoo, FreeBSD, and so on, there's more work to do for wireless printing because you need to properly set up network connectivity, install CUPS, and check if your printer requires drivers that aren't already in the kernel. On any distro with a preconfigured GUI like Mint, Fedora, MX, antiX, even Artix or Omarchy, you don't necessarily need to all of that. The most you'd really need to do is double check that your specific printer model doesn't require additional drivers. I'm sure the technology for wireless printing's existed for longer than I think it does, but I think the capacity for it existed since the early 2010s.

Also your fonts being fucked is fairly normal on minimal distros like Arch or FOSS-forward distros like Fedora. By default, a lot of the most common fonts we see on the internet and IRL are patented or otherwise non-free. You'd need to either manually copy font files over from a Windows installation you have running or hunt down the MS core fonts package for your distro so that you can at least pretend that your system ain't totally bonehurt.

It isn’t, to my knowledge nobody serious actually uses blender. It might be a good pice of software but it doesn’t come with a warranty or 24 support and other things professional shops care about. That’s another barrier for FOSS software unfortunately.

Not entirely. There's also the flip side to consider, where libre software forms the basis for commercial non-free software. The norm is to shit on permissive licenses because the MIT license allowed Intel to hijack MINIX and use it as the OS that powers Management Engine, and Andrew Tanenbaum couldn't do jack fucking shit about it. That norm exists for good reason, but we should also consider how not every permissively licensed project would end up like MINIX.

FreeBSD is used by Sony to power the PS4 and PS5's operating systems, WhatsApp uses it on their servers, as does Netflix and many others. These companies not only use FreeBSD, but they also submit their own patches and modifications back upstream that later get integrated. FreeBSD is not "base system + X11/Wayland + KDE/XFCE/whatever + all the shit I use like a normal human being;" it's literally the base system alone. Y'know, that text console you boot into after installing it? Sending bug fixes, patches, modifications, and so on back to FreeBSD doesn't hurt those companies because all the secret sauce is in the upper layers, where the UX lies. The opposite is true; it behooves them to cooperate with upstream.

Also, lest we forget, Wine is free software that's not permissively licensed, but its copyleft is weak enough to allow for integration into proprietary programs. Codeweavers literally has most of the development team on Wine under their payroll, they take the Wine source code, and add some UX tooling and commercial support that they keep behind a proprietary wall. The guys who run pfSense and OPNSense have their own private companies, where they take pfSense/OPNSense, put it onto some network box that they sell on their website, and that network box has proprietary tooling and UX shit that ain't free to the world.

Copyleft is ideal, but we shouldn't let Tanenbaum's misfortunes override pragmatic reality on the ground. Sometimes, a weak copyleft or permissive licensing is a good thing.

Granted for Linux RHEL exists for this exact reason and what our wonderful community fucking despises them for it.

I resent that! I don't despise Red Hat for monetising open source; that's like the biggest good thing they've ever done. I resent Red Hat as a company for embracing, extending, and extinguishing competing software projects in favour of making their own tooling dominant within the Linux ecosystem long before their acquisition by IBM. See all the systemd sperging, how Red Hat actively sabotaged Xorg development to push Wayland, the myriad problems that Freedesktop has, the circumstances that led to XLibre's creation, and so on. Canonical has my ire too, but to that end? I prefer the Englishman from Cape Town over the faceless stooges at Red Hat and their faceless masters at IBM.
 
Printing on Linux actually benefitted from Apple of all companies. I believe it was Apple that created CUPS (the Common Unix Printing System), released it as open-source software, and it basically became bog-standard on everything from OSX to Linux to BSD and even whatever proprietary Unices that still technically exist like AIX, HP-UX, and Solaris. USB printing was always seamless for me, even like 14-15 years ago when I first started fiddling with Linux. I never had a network printer until recently, and as far as I could tell, the process was seamless. Printed out test pages just fine, scanner recognised, all that good stuff. I should point out that this was with the aid of GUI tools in MATE, Cinnamon, GNOME Shell, and Plasma 5.
Not only that, Apple also fixed printer drivers with AirPrint. Basically, they required that all network printers be compatible with one generic driver, and the result is we can just connect the printer to the network and then print from just about any device without needing to install a separate proprietary driver on each device.
Apple are the greatest thing to ever happen to printers. Printers are still awful, of course, but now it's because of legitimate issues like a printer is a mechanical device that can literally get itself jammed, and not just because HP wants to install their malware in the guise of a required driver so they can advertise their own overpriced toner packs whenever you start running low.
 
Is there a trick to making games actually run on Linux? I've tried 6 so far and none of them work except for Half Life 1.
Yes I'm using proton. It just downloads automatically, right?

Yes, Proton should download automatically when you install games from Steam. A better question to ask is what games are you trying to run, and on what kind of hardware? i.e. are you trying to run Destiny 2 on Linux without realising that games like Valorant, Apex Legends, Destiny 2, Battlefield 6, and so on all require kernel-level anticheat and whose developers are hostile to Linux?

ProtonDB should give you a rough idea of whether or not your game will run on Linux.[/url]
 
Last edited:
Yes, Proton should download automatically when you install games from Steam. A better question to ask is what games are you trying to run, and on what kind of hardware? i.e. are you trying to run Destiny 2 on Linux without realising that games like Valorant, Apex Legends, Destiny 2, Battlefield 6, and so on all require kernel-level anticheat and whose developers are hostile to Linux?

should give you a rough idea of whether or not your game will run on Linux.
I tried to run tf2 and it wouldn't download all the way it just kept saying "corrupt update files" then I tried CS2 and it did something similar. And finally I tried Ready or Not which just didn't launch.
 
Back
Top Bottom