Linux on the Desktop


Firefox stopped supporting ALSA. It was as if dozens of voices suddenly cried out in terror and were suddenly silenced.

I feel like tablets, macbooks, and Windows 10 combined have made Linux on the desktop entirely irrelevant for anything but development environments or highly specialized nerd crap.


I, for one, am glad to see it go. Linux on Desktop has been, primarily, the land of pretentious nerdery for a while now.

“Oh, you use WINDOWS? Too bad you had to pay the Micro$oft tax! I use Klimbintu, a TRULY OPEN Debian OS. Not like UBUNTU. I hand-made my drivers. I can’t use firewire or the built in Wifi on my laptop, but at least I know it’s only doing what I WANT it to do. Look at me! See my command line? I use pdksh, not sh like those plebian casuals. I type in DVORAK inverted.”

Well, fuck you, you fucking hipster.


There’s freedom on the one end and walled garden on the other.

If linux disappeared, everything would be awesome provided you used everything the way the company who made it dictated and doing otherwise would be impossible.

If Windows/MacOS disappeared everything would be free and nobody but nerds would be able to even use a computer.

Both worlds are shitty. As the walled gardens allow you some degree of freedom, the need for Linux decreases. If they start more locking down as the norm it’ll come back. Much like the tide it’ll be back and then leave again.

Rename this thread: “Tides are a thing”


But Black Dynamite!

(Posted from Debian GNU/Linux 8 (jessie).)


It was viable, to a degree, back in the mid 2000s if you didn’t do PC gaming. We produced GeekNights entirely in Linux, and it was my main desktop during that period.

But back then, the OSS clones of “real” software were closer in functionality than they are today. Audacity hasn’t gotten much better than it was back then, all the Office clones are stuck with 1999 feature sets, Rezound hasn’t been updated since 2008…

When Adobe took over CoolEdit, they didn’t add anything useful to it for a long time. But then they started adding to it. Audition is now so far beyond even the best OSS audio software that there’s zero comparison. Photoshop isn’t just a couple generations ahead of the GiMP: it’s in an entirely different class of tools.

OSS on the desktop thrived when it could easily replicate the features of commonly used software. Clone enough of Word, and people have enough to get by. Now there’s no hope catching up

Windows is far from a walled garden. Call me the day Microsoft actually prevents me from installing software on my desktop.


I agree, this is part of why linux is the least necessary it’s been in a long time.


But only for desktops. Servers run Linux or you’re doing something terribly, terribly wrong.


The idea of artisanal drivers made me chuckle.


Linux on a desktop is terrific because

a) It’s free
b) All the important software is open source, you can just install it from a repo. Like an app store where everything is free and trustworthy and you never have to go elsewhere.
3) No ads, no spying on you, no bullshit, constant updates.
4) The fastest and most stable.
5) It’s the best development environment.
6) Lots of choices. You can choose what desktop you like best, etc.

Linux on the desktop fails for two reasons.

  1. Drivers. Sure. Just about all the common PC hardware works on Linux. There’s even that guy who made drivers for every webcam ever. The problem is that that all the fiddly features won’t work. Sure, your webcam shows video, but good luck getting the Logitech autofocus face detection thing working. Sure, your sound card plays sound, good luck with surround sound.

Consumer peripherals are the only things that fail here. There is a chicken and egg problem. Manufacturers aren’t going to make Linux drivers unless it reaches critical mass. It won’t reach critical mass if they don’t make drivers. Even if it does reach critical mass, many companies don’t want to include free open source drivers in the kernel itself. That’s how you get NVidia drivers shenanigans.

Apple solves this problem by not letting you choose your hardware. Windows solves it by having support from an entire industry.

  1. The benefit of choice is also a curse. Part of the reason we have operating systems in the first place is so you can write software once and create a binary that will run on many different machines that are just similar enough. Remember when you had to choose your sound card from a list when installing DOOM? Nowadays you have things like DirectX abstracting away that nonsense, so you can write one game that runs on every Windows machine with every sound card.

On Linux, those sorts of things are user choices. A game that runs on Debian might not run on Fedora because there are subsystems that are different. It’s not a problem for a simple application, but anything fancy like a game or a web browser is a large effort to make it work everywhere. The distributions themselves put a lot of work into making working builds from source distributions. They can’t do that for Counter-Strike, so Valve has to do it themselves, or only support a limited set of Linuxes.

For Linux to succeed on desktop they have to standardize and lock in the userland APIs and get broad support from the industry. The window where that could have happened has long since closed. The desktop Linux is now only useful as a platform for developers. It is also OK for a computer that doesn’t need to do much (grandma’s PC) where you want to save money on a Windows license.

All of this is 10+ year old news. When Windows 7 was released I think that was the final final nail in the coffin for desktop Linux being a thing. There really is no reason to discuss it anymore.


Talk to some of my clients, I swear to you, idiots at some giant companies insist on using windows servers. We had to hire additional devs to keep our windows software in pace with our actual product.


They’re doing something terribly terribly wrong. The only time you need a Windows server is to provide actual Microsoft services to your employees. For example, you might have an Active Directory or run Exchange. For all the standard services like DNS (BIND), DHCP, HTTP, NTP, etc. the best server applications run on *NIX and the best *NIX is almost always some flavor of Linux.


There exist chat bots that talk to millions of users daily that are running on windows server 2013. There is exactly 0 reason why they don’t run on linux, I support linux for other much better clients.

Just having this job caused me to sell several stocks in my portfolio on sheer incompetence I’ve witnessed having worked with their ops teams.


If you believe some of the more scary future predictions, where the bread becomes scarce and the circuses rare, there may be a reason to continue to discuss this.


How do you even get a service to run on Windows like that? Was it built with Visual Studio? Does it appear in the “Services” section of Windows with the yellow gear icon next to it?

In *NIX you write a program and then have some sort of init script to launch it depending on your init system. Even on OSX you use launchd. It sucks, but it mostly works if you figure it out.

But on Windows? It’s a fucking mystery. I write a simple HTTP server, let’s say. Do I just make an exe, put a shortcut to it it in the Startup folder? Does it appear on the desktop or notification area while it runs? That’s way janky and not server-like. It should be completely hidden and silently running perfectly in the background.

Also, Windows is crazy unstable, relatively speaking. How do they deal with the lack of uptime?


It does appear in services, and there’s a property in there where it can be manual or automatic start.

It is invisible outside of the service and process (task in windows) it creates.

As to how they deal with the lack of uptime, well it’s a simple cost benefit thing. They have to pay for uptime using the headaches of sysadmins. For the low low price of about 1 headache/day the uptime can be nearly perfect.

(they use apache+tomcat to launch the exe)


Even back when I was a teenager in the mid 2000s and was just futzing around with Linux to see what the fuss was about I was left cold. The entire time I was thinking “I can see how this is very powerful but I just want to play my games and type my school papers and pirate movies.” I wanted it to just work without a lot of rigamarole, and I reached the same conclusion that it was a geek hipster stick-it-to-the-man thing outside of specialized use cases.


To Linux’s credit, it will put up with weirder situations than Windows will. 4-5 years back, my graphics card totally barfed and Windows simply would not run with it having drivers installed. No drivers, though, and it would work fine. I installed linux on another drive and it WOULD use the graphics card. However, trying to use any of my audio equipment or edit video or do anything that wasn’t basic office tasks was a pain and a half. I had to use janky firewire drivers that failed every 20 mintues and then I had to…
Get JACK to play nice. I got it working after a week and a half of fiddling, but it just wasn’t worth it. I finally just gave up and borrowed some money to go buy a graphics card. And BAM, windows was BACK and everything was AWESOME again.


I use it daily with no intentions to go back as my main OS, for the positives Scott posted above. It definitely has it’s issues and it can be fiddly but it suits my needs and runs well on older hardware. I won’t argue that the OSS media editors are just as good but they generally do anything I’d need since I’m not a professional. Even though they’re getting better about it, games are still an issue which is why I bought a dedicated Windows PC for that. Hopefully I can eventually just use a Windows VM for the exclusives and have a Videocard passthrough. I’ve had some success with Wine but Steam in Wine just seems like too much hassle.


[quote=“ninjarabbi, post:18, topic:574”]
I won’t argue that the OSS media editors are just as good but they generally do anything I’d need since I’m not a professional.[/quote]
Oh, god, no. They’re all basically garbage. Any NLE for video or audio that is actually worth even trying is closed source. The best open source editor for audio I could even try was Ardour and it’s basically GIMP but for audio. It has a lot of the right features, but it definitely feels like it was designed by a programmer rather than an audio engineer or editor.


I would still and do still use it for any machine that doesn’t need to play games. However my niche use case is programming so it shines incredibly well under those circumstances.

I also agree with the positives and negatives listed by Scott.

Having community vetted code repositories for programs is great and allows for quick fixes if anything goes wrong.

Many external peripherals now just work on Linux because they have on board drivers and register themselves as generic devices. For example my headphone amp just plugs in by USB and it starts working however on Windows it needs a driver to function.

I actually like the Gnome 3 desktop and all the features you can customise with it. If I don’t like it anymore or just want to try out something new I can get a new desktop environment.

All the negatives do bring it down for all other purposes unluckily, however if it’s just general computing, you can get away with 99% of use cases fine.

I don’t like the open source zeal that is often built into distributions but I can understand why it was important at some time.

Linux on the desktop does live on through Chromebooks however, I’m considering getting one of those later in the year if Google implement the Android app function well. Otherwise I’ll just end up getting Microsoft Surface Pro.

Also remember you have a Linux subsystem on your Windows 10 lol.