The Linux Thread - The Autist's OS of Choice

  • Want to keep track of this thread?
    Accounts can bookmark posts, watch threads for updates, and jump back to where you stopped reading.
    Create account
This is also the selling point of Linux desktop distros (particularly Mint).
There's a kind of reverse Dunning-Kruger effect that people like you experience, that the average skill and intuition of the population is higher due to your own skill and familiarity.

There's a reason I harp on about command prompt and GUI. The average struggles just a bit with a gui with colored buttons, and you expect them to learn command syntax? They see it as something dangerous that they aren't supposed to be doing, and will just copy a command from an existing forum post outright even if that command bricks their computer.
With all that being said, in the hands of an experienced user, yes, Linux really is that good despite what distro you choose. But there simply don't exist that many experienced users.
You know, this is a disconnect I didn't really think about between myself as a windows user and the average linux user: It's not just that it seems hard to adjust to, it flat out seems like a downgrade in the way linux does things versus the way windows does things for the average user. I've never heard of a package manager, and I don't understand why I would need one over just installing executables off the internet. Ricing? What style are you adding that's so cool it can't be added by a wallpaper, or if you are feeling really frisky by wallpaper engine? Does it matter that it's easier to solve hardware issues by using terminal commands over the labyrinth-like windows submmenus, when it's going to be the same choice of following the first result of a google search and you would have pictures telling you what to click over commands they don't understand?

It's not just a lack of experience, it's a complete non-understanding of what exactly makes the linux way of doing things any better, or not just flat out worse than what they have already been doing. Especially when you tell them it's also going to cost any of the extra proprietary software and hardware they might want to use. It doesn't seem like using a gas stove over a fire place, it seems like using a thimble to slowly put water from one pot to another instead of a pitcher.
 
I recently discovered this oddity while snooping about the GRUB settings, apparently you can have GRUB play a tune with the PC speaker before the menu loads if you have it set on timeout/menu. The tune is entirely custom, one channel, chiptune and is set up entirely in GRUB's config (GRUB_INIT_TUNE) with this note pattern: ''tempo freq duration [freq duration freq duration...]'' I haven't heard anyone talk about this before, and probably because it's pointless. I guess it's some legacy feature from back when sound was more a roll of the dice on Linux? There's a few people who have shared some jingles and even entire songs. - Be aware that you can't skip it or do anything while it plays so if you put in some long ass bullshit you'll have to sit through the entire thing on full blast every time you boot.

I went with the Coin SFX (600 988 1 1319 4) myself that I just stole from the website, you can test + compose on the same website if you have the skills to do so.
 
I recently discovered this oddity while snooping about the GRUB settings, apparently you can have GRUB play a tune with the PC speaker before the menu loads if you have it set on timeout/menu. The tune is entirely custom, one channel, chiptune and is set up entirely in GRUB's config (GRUB_INIT_TUNE) with this note pattern: ''tempo freq duration [freq duration freq duration...]'' I haven't heard anyone talk about this before, and probably because it's pointless. I guess it's some legacy feature from back when sound was more a roll of the dice on Linux? There's a few people who have shared some jingles and even entire songs. - Be aware that you can't skip it or do anything while it plays so if you put in some long ass bullshit you'll have to sit through the entire thing on full blast every time you boot.
so you can make it do the macOS startup sound?
 
I'm switching to it whenever I find time to sit down and play musical partitions: In news that that will surprise nobody, Manjaro has shat all over my dependencies again and I'm out of fucks to give and time to waste fixing the damn thing. @Croco if you want to spend all day staring at a terminal prompt you should give Manjaro a go.
The idea is not to waste time staring at terminal. It's that the cli is scriptable. You can make a bash alias, +x shell file, whatever. And reuse the scripts. $ mv * ~/Folder is easier than holding a shift key and dragging dropping all these files. Larry Wall calls it laziness impatience and hubris.
Manjaro is shit too. Arch is not hard anymore to install, and you can just use XFCE. No need to rice anything if you don't want. It's all just a bootstrap to load Emacs for me.
 
The idea is not to waste time staring at terminal.
I disagree.
1747886848821.webp
 
I recently discovered this oddity while snooping about the GRUB settings, apparently you can have GRUB play a tune with the PC speaker before the menu loads if you have it set on timeout/menu. The tune is entirely custom, one channel, chiptune and is set up entirely in GRUB's config (GRUB_INIT_TUNE) with this note pattern: ''tempo freq duration [freq duration freq duration...]'' I haven't heard anyone talk about this before, and probably because it's pointless. I guess it's some legacy feature from back when sound was more a roll of the dice on Linux? There's a few people who have shared some jingles and even entire songs. - Be aware that you can't skip it or do anything while it plays so if you put in some long ass bullshit you'll have to sit through the entire thing on full blast every time you boot.

I went with the Coin SFX (600 988 1 1319 4) myself that I just stole from the website, you can test + compose on the same website if you have the skills to do so.

Yeah this is some arcane knowledge. My dad set up his PC to do this before boot in 90s. That was ricing back then.

@Aidan Cbonsai is cool too. I haven't tried out aquarium yet.

:story:
 
I've never heard of a package manager, and I don't understand why I would need one over just installing executables off the internet.
That particular point seems pretty easy to explain to someone. A package manager means instead of having to look up individual sites to download programs from you can just open the terminal and type "install X" and X will be on your computer, and instead of needing to individually keep all these programs updated you can just periodically open terminal and type "update" and everything is updated. Pretty convenient.

Of course, in reality this is overselling the average package manager, since unless you're running Arch the repository your distro is using likely doesn't come with every program you'll want or in its most updated state. So you'll actually still likely be going to websites and github pages and downloading precompiled binaries and appimages, pretty much the same as you'd be doing on Windows. Which is something I don't really see people mention, but maybe that's because so many people will just use Flatpaks for everything instead. But personally I've never really seen the point in Flatpaks in most cases, and they've even worked worse for me than other options a few times.
 
It's hard to explain the purpose of a package manager to someone used to the windows way of doing things.

To them, a program is just a blob you download and click and it does things. How does it do things? Doesn't matter, it just does, and if they can't do those things on another operating system it's because it's a bad operating system that can't even run programs.

You could start by explaining that programs all must interact with the system to make use of its services for such basic functionality as I/O, and the precise mechanism for this varies by OS, cpu architecture, and ABI. Not only does the mechanism for using its services vary, the behavior of those services and the interface to them also varies. Usually you can abstract away many of the differences of mechanism by using a "standard" library that handles these variations transparently, but the mechanism for actually using that library itself is also going to vary by OS, cpu architecture, and ABI, so in practice you need to use some sort of mechanism for going from portable code to nonportable code, usually a compiler of some sort.

So when you download a program and click it, you actually have to download the correct one for your OS, cpu architecture, and ABI. You just don't think about it because Windows has one OS, one cpu architecture, and these days effectively one ABI. Of course, a more general solution is to just get the source code and convert the portable code to nonportable code yourself - that will automatically Just Work for any platform with a suitable compiler.

But to the windows user this is actually less generic, because it's harder to get ahold of portable source code than nonportable code, because of the shitty culture.

You could continue to explain that actually, programs can be parameterized at compile-time such that a different nonportable program will be produced depending on how you configure it. This allows you to do things like select what defaults the program will have, what its dependencies will be -

But wait, what's a dependency? A windows user has never heard of a dependency, and is baffled by the thought that someone would distribute an incomplete program - a program that doesn't even work! You could bring up the various benefits of reusing code libraries across multiple programs - modularity, security fixes, disk space, etc - and at best you might get a "well then why don't they just install the library when you install the program?" You could bluster on about DLL hell, about missing security fixes ("then just automatically install the latest version!"), about incompatible variants or versions, etc, but I'm sure it would sound like gobbledygook to them.

Perhaps you would give up on that for the moment and return to describing compile-time configuration, like choosing where the program being compiled will be installed. At this, a faint glimmer of recognition arises: maybe, one time, while mindlessly clicking through an installer, they saw an option for where to install the program files to. But wait, why would it need to know this at compile-time? Because that's how it knows where to find its own files, and references to them may be included in the nonportable compiled code. It also determines at compile-time where to find the files of its dependencies. At this their brow may furrow a bit, as they struggle to make the connection between being able to choose where a program goes and having to figure out where another program went.

Suddenly, a flash of inspiration: "you're wrong, windows programs are portable! I once ran a program straight from a flash drive!" Of course, this was only possible because its dependencies were all in the One True Place (windows programs are always in the One True Place), or statically-linked. In other words, windows programs are only portable because the programs they depend on aren't portable. But this explanation cannot shake his Lived Experience: windows programs are portable!

At some point you would realize that if you can't even convince the windows user that there is a point to compiling a program, it's quite hard to explain using a package manager. Perhaps you would fall back to a social argument: it's dangerous to just go around running random executables off the internet, and widely-used package repositories afford some safety in numbers. But of course, this immediately exposes you as the arch-nemesis of the windows user, the Mac user (there are after all only two operating systems, windows and Mac!), with your pathetic freedom-hating walled garden!

Eventually it'll become too much. You'll smile, nod, and say "yeah, package managers are a stupid linux thing we use because our operating system can't even run programs haha". It's better to let them enjoy their one OS developed by one company on their one architecture with their one ABI where programs all bundle their own copies of libraries and use one system library always in one place, and users limit themselves to one configuration of a program that someone else compiled. Let them think "I am free".

You think this is satire. Try looking on NuGet sometime. It's incredible. Even to windows people using "package managers", a package is just something you install with the equivalent of 'cp -r'. "Oh yeah, we use a package manager, our MSBuild files download an entire compiler as a binary blob from NuGet to a fixed location and then run it to produce our output dll, then we download all our dependencies as binary blobs from NuGet and include them alongside it in the zip file".

And the worst part is, this kind of cancer is spreading into FOSS via language package managers. It's absurd how many pypi packages can't be built without an internet connection, and don't even try to offer a source package. The open secret about flatpaks and snaps and whatever other bullshit is that one of the reasons they're being used more and more is that developers - especially "open source" developers - are steadily getting worse and worse at making packages that are actually able to be built from source with any kind of quality control. So many packages can't make it into debian or any other distro with standards, and the developers stubbornly refuse to fix their shit, so all that's left is binary packages in the form of flatpaks and whatnot.

Don't even get me started on the insanity of evaluating licensing information for a package that bundles binary blobs of unknown provenance in its source repository.
 
I briefly explored it years ago and said "lol no" as a daily driver. Great on some servers, though. Try it and see? There's definitely less support for it so it depends on you and your hardware I guess.
I'm not saying don't install it. Just go into it with the right expectations. You will have to do more setting up than any Linux distro I have tried at least.
Thanks for the heads-up.
I have used OpenBSD on a laptop before and it worked well. I used dwm on it just like on linux and it worked the same way. Wifi was actually easier to setup on openbsd than on linux, it's simpler in design. Same for audio. I might switch back to openbsd now that linux is being taken over by flatpak and wayland cancer.
I was looking at OpenBSD as well, however, it does seem to be geared for a specific use, and FreeBSD more "generic", so I'll start there and see. And yes flatpak is cancer, Wayland I'm not sure - yet.
Honestly, what I'd like is a system that's at least trying to be cohesive, not just random stuff slapped together. I know that's anathema to UNIX in general, but there are degrees to that.
 
Honestly, what I'd like is a system that's at least trying to be cohesive, not just random stuff slapped together. I know that's anathema to UNIX in general, but there are degrees to that.
Slackware is honestly a good distribution if you want something that feels more like one cohesive thing.
A bit for advanced users but if you are looking at a BSD its a really good "halfway point", very old school (which BSD is) in the way it does things, but with modern Linux benefits as well.
Not quite as scary to make the leap to if you want to try out a bit more of an old school way in a typical Linux environment.
For a bit of fun I would still check out FreeBSD. Its a pretty terrible desktop OS as others have said but you can make it work (with pain and suffering). Don't expect everything on a laptop to be functional.
I do remember Wifi on FreeBSD being absolutely horrendous but I do know Nvidia works way better on BSD then Linux, so your mileage will vary based on your hardware.
But its also been a few years since I seriously looked at a BSD so maybe things have changed, I remember hearing Wayland somewhat works on FreeBSD now so maybe more modern stuff will work too.
 
Congrats, dumbest comment I've read on this website.
What exactly is dumb about it? Most consumer-oriented linux flavours mount drives on /media (or /run/media because poettering, as you pointed out), MacOS mounts drives on /volumes/, and Windows mounts them as drive letters, but the thing that actually matters in the GUI world most users inhabit is whether they can click on an icon and get their stuff. They don't care about the implementation behind the scenes. They want their stuff. Even if they're advanced enough to be using a NAS, they don't care that it's mounted on /volume1 or whatever, because they're using the icon that has their stuff inside it.

I recently discovered this oddity while snooping about the GRUB settings, apparently you can have GRUB play a tune with the PC speaker before the menu loads if you have it set on timeout/menu.
I was looking up this exact feature just yesterday. Talk about synchronicity.

Honestly, what I'd like is a system that's at least trying to be cohesive, not just random stuff slapped together. I know that's anathema to UNIX in general, but there are degrees to that.
MX works for this. It's the most coherently put together distro I've seen, at any rate.
 
What exactly is dumb about it?
Who exactly was talking about a retarded normie strawman? Not me.
And I pointed out how automount sometimes just fails, either down to the distro not even including it or it failing for bizarre reasons.

Both of these things bother me immensely based on my own usage across the past 20 years. It makes basic file tasks with external drives annoying.

So yeah it was pretty fucking dumb.
 
I haven't heard anyone talk about this before
Probably because PC speakers are now a rarity in most systems, especially after Windows Vista/7 removed driver support for it. You were only left with some third party drivers that could re-enable it, just about every computer had integrated HD audio and the PC speaker only made sense for debugging. I know that back when I managed to get it working I made an AutoHotkey script that would play L'Amour Toujours and the Bad To The Bone riff. Made another one that would play the melody, show a gag notification window, then after closing it it would play it again. Shame that Microsoft screwed over the fun by killing off the old Beep driver.

Not to mention my motherboard is fucky when it comes to supporting the PC speaker, but now at least it makes a happy little chirp once it's done POSTing and as long as I don't touch the motherboard it should remain that way. Might try to get a proper speaker off of Aliexpress and resoldering the plug from the small piezo shitstain I'm using right now to give it some proper beefy beeps.
I guess it's some legacy feature from back when sound was more a roll of the dice on Linux?
No not really, the PC speaker is an extremely basic way of making noise that existed ever since the first IBM PC so it doesn't need any complicated audio drivers to work, which is why GRUB had the option to have a fun custom jingle play once it loaded up. Remember that for a proper audio jingle in GRUB, you'd need to load both the sound file and a rudimentary audio driver stack into the EFI partition which is already tight on available space, and doing all that shit for that extra rice is just pointless. A PC speaker melody on the other hand is super cheap to implement.
so you can make it do the macOS startup sound?
See above, though IIRC there were ways to make the PC speaker play PCM samples akin to how C64's SID chip could do it, but then again, wouldn't be the same. The macOS startup sound is unique to Mac hardware since Apple controls it from top to bottom, so they have their own little implementation of the sound stack where it will play that sound once it finishes booting, and unlike Windows/Linux, this is way lower level, hardcoded in the motherboard. In the Windows/Linux/"x86 IBM-PC compatible" realm you only get that POST beep at that level.
I've never heard of a package manager, and I don't understand why I would need one over just installing executables off the internet.
Try using winget at some point, it's super comfy to be able to install, update and remove software from the command line. And once you delve a bit deeper into it you can make PowerShell scripts to autoinstall shit in batch, so instead of manually looking up all the installers, downloading and running them, winget will automate it for you, and all you have to do is run a single script.

Here's a good example. You have a fresh Windows 11 install. You want to install Firefox. What's more convenient?
a) opening up Edge, going to firefox.com, manually downloading the installer then running it
or
b) Hit Win+R, type wt, hit Enter, type winget install firefox and hit Enter again?
The latter is obviously way faster and more convenient. You still have the option to do it the old fashioned way, but do you really want to if there's a better, faster way available to you?

Fun fact: package managers on Linux work more or less the same, they're meant to pull software files from the Internet, extract them into all the right places and add all the necessary configurations to the system. Windows installers do the same, and winget is essentially an automated way of pulling and running those installers. Bit more convoluted and bodgy than on Linux but the end result is the same.

Remember that the terminal is not exclusive to Linux. Windows started off as an overlay for MS-DOS which was a command line OS, and it still has a command line that is incredibly powerful once you wrap your head around it. The difference is that Linux is a CLI OS with GUI elements added on top of it so in the end you will be more reliant on it. Windows NT was built from scratch to be GUI based so for the most part you'll be navigating legacy GUI hell instead of dealing with the command line.
 
No not really, the PC speaker is an extremely basic way of making noise that existed ever since the first IBM PC so it doesn't need any complicated audio drivers to work, which is why GRUB had the option to have a fun custom jingle play once it loaded up. Remember that for a proper audio jingle in GRUB, you'd need to load both the sound file and a rudimentary audio driver stack into the EFI partition which is already tight on available space, and doing all that shit for that extra rice is just pointless. A PC speaker melody on the other hand is super cheap to implement.
Ah, so I guess it's more for something like a data center rack, where you can hear the tune and know the machine has reached the boot loader instead of being stuck in POST or something? That's neat I guess. I suppose that's why it works on my old laptop too that still has this feature.
 
It's probably a skill issue
It most likely is since PowerShell is object oriented, unlike Batch or Bash which is strictly string oriented. For example, to pull out the uptime, you pull out a date object which then you can format to pull out individual values. Or when pulling out the local IP address you pull out an object with all the information of the interface, then you pull out the IP value. It's not better or worse, just different with it's own pros and cons.

I do agree that Bash is more intuitive than Batch. Writing a Batch script feels like you're writing it wrong on purpose while Bash has a sense of consistency. My yt-dlp script for Batch feels like a completely different beast after rewriting it in PowerShell. Instead of having to manually declare menu positions and fucking GOTO commands I just add a new entry to a JSON file and I don't have to touch the script, it just works.

By the way, if you want to inflict it upon yourself, PowerShell 7 is multiplatform and can run on Linux. 99% of Linux users despise everything Microsoft on principle anyways, but I guess the handful of sysadmins that like PowerShell appreciate it.
Ah, so I guess it's more for something like a data center rack, where you can hear the tune and know the machine has reached the boot loader instead of being stuck in POST or something? That's neat I guess. I suppose that's why it works on my old laptop too that still has this feature.
Just a relict of a bygone era really. Computer cases used to have those mounted by default and the SPEAKER connector was a part of the standard front I/O.
1747925366248.webp
Later on the full-sized speakers got replaced by these small piezo dinguses just so that you could hear the beep codes when the motherboard shits itself.
1747925377548.webp
It's good to keep a few of these on hand for diagnosis, the motherboard beeps helped me diagnose that the issue was RAM being fucky and that it had to be taken out and put back in again. I'd suggest to get them from Aliexpress since they cost pennies, but most users here are from the US so lol.
 
Back
Top Bottom