This is accurate. The Linux desktop is an experimental success that persists with major flaws. If it wasn't, 99% of us would be using it instead of Windows and macOS.
Absolutely zero requirement or expectation to be technical is the biggest one IMO. For example you don't have to know or care what an operating system is or does to use macos, which is a lot less congnitive load. And some Linux based OSes solved this too, like ChromeOS and Android. One of the big value props for specialization in civilization is that you can just worry about your own thing - like idk how to fix a broken down NYC subway train and I don't really care as long as someone does. And Linux pretty handily wins in server land where the users have specialized into being technical and that's their job.
I'm pretty sure a non-technical user can use Ubuntu or something, from start to finish, without ever touching a terminal.
To me, the main issue is that many preferred softwares are not offered on Linux because of its small market share. The market share doesn't grow because preferred softwares aren't there. It's a circular problem.
That issue is improving all the time though as new softwares tend to target Linux natively (ex. Discord, Zoom, Slack). Furthermore, the compatibility layers are getting better for non-native programs (ex. Steam deck).
I really don't see a universe where Linux stops taking over the world, including the desktop. Adobe and Excel probably can't hold Windows up forever.
But right now they'd have to install ubuntu or know to buy a machine with ubuntu pre-installed. Compare to a chromebook, where it already has a linux distro, chromeOS, installed. And at this point I don't think software compatibility (except for gaming) is the biggest issue since thanks to electron/browsers most things should just work.
And for zoom, in addition to asking you to select your 'linux type,' has this little gem: "Zoom's rpm packages are signed with a GPG key. Please run "rpm --import package-signing-key.pub" to import the key in case package management utility asks for a missing public key" - though I guess probably the auto-installer thing would run when you opened a zoom link and skip all that. I doubt it's the only time a linux user is assumed to know what a terminal is or expected to know how to use it.
On the other hand - for linux - WSL is pretty cool. That's definitely a lot of linux on a lot of people's desktops, whether or not they use it!
I hereby officially withdraw my comment about windows users not needing to understand technology after the everything with crowdstrike and the fix instructions involving booting into safe mode and deleting files with system32
I think anyone who can find the right window and fix an install using just the UI has put a lot more work in than someone who could copy complete working answers from stack overflow and now GPT. Of course they have to train at it over and over since they can't use shell history so they no longer think it is a real skill to find these absurd boxes.
I've found that it's pretty good at this point for the inlaw/grandma crowd that mostly just checks email and Facebook. As long as you (someone that isn't them) installs and configures it for them. Pop is my current recommendation and it's working well for a few people I've setup. My grandmother ran Ubuntu for close to the last decade of her life, I only had to update now and then, and I setup WINE for the two games she wanted installed.
This isn't too much different from how a lot of people use Windows and Mac, many wil have their kids figure out how to setup an application for them. It depends, of course. A lot of professional apps just don't have suitable equivalents for Linux and WINE doesn't do it for them.
I'll put it this way, my non technical friend is tried of Windows.
She doesn't play video games or need much space. An M2 MacBook Air is 800$. If anything goes wrong, just take it to the Apple store. Even out of warranty if it's something simple like you broke a setting, they'll fix it for you.
It'll be much easier to just buy a Mac. I personally don't want to buy a new Mac because I want to have a 4 TB SSD and to get that in a MacBook you have to spend between 3k and 4k. Vs being able to upgrade a cheaper Windows laptop for 200$.
I personally dual boot Linux Mint and Windows. Linux is much better for getting things done.
Networking is one thing that can get especially weird. And command line might not work... You have too many ways of doing same thing and they kinda work together until they do not and then well nothing saves you apart from cleaning up everything and starting again...
It hasn't been like this for a very long time, but "You need to compile the kernel in order to get it to play sound" is still an incredibly effective attack because it puts advocates on their heels trying to defend it as baseline usable.
When you think about its main competition it's a little silly to keep framing it this way, but again, it's extremely effective so it doesn't really matter if it's right.
Kernels too sometimes stop working. Am currently testing latest stable (caveat with some patches). Nothing major has changed as far as I can tell from my sources but the newest release (6.9.9) now has an unexplainable panic on boot.
No need, nor expectation to open a Terminal/Cmd/Powershell and start writing cryptic commands. Most are go here and set this slider on/off. More advanced are open regedit and write this variable on this path to 0 and restart. Or Open group policy editor and set this setting on this path to Disabled.
AppImage are perhaps that only thing working
OOTB like windows/Mac programs,
and other container formats don't come close.
Everything else is dependent on right configuration,
correct versioning and lots of
"magic" settings that need to be
set to make something functional.
Depending on correct glibc, dependencies
being installed with the right settings,
environment variables,makefiles
being compatible with what your system
has,having the right kernel version,
etc.
> Depending on correct glibc, dependencies being installed with the right settings, environment variables,makefiles being compatible with what your system has,having the right kernel version, etc.
Apart from having a bingo at what dependencies you bundle, AppImage makes no effort to actually solve this, where as Flatpak and Snap do. The reason glibc, environmental variables, dependencies is important is because the Linux desktop is not a single distribution etc, nor is there one arbiter keeping ABI stability.
You venture out onto a distribution that is not what the AppImage was built on and you see what a wild ride it is of if things are working or not. Incompatible curl versions, glibc, mesa, etc.
AppImage pretends to work like .exe on Windows, or .app on macOS, but it doesn't share anything similar with them at all. Both Windows and macOS have stable userspaces where things don't change, Linux on the other hand...
For refence, Flatpak solves this via freedesktop runtimes and Snap solves this by installing Ubuntu containers.
My current gut idea is that all of these discussions are woefully incomplete without "following the money" (and or power?)
For example, I don't think you can adequately explain GNOME's "popularity" without taking into account inertia from Red Hat and other companies, given how much it arguably sucks, and un-arguably doesn't play nice with other stuff.
I guess the least bias statisic for the popularity of DEs would be Arch's pkgstats [1] and considering I would classify arch more of a tinkerer distro I would still say this is biased twards KDE from the type of user, but that is just my specualtion.
Things are popular when they are only or first option that get automatically installed. Defaults over power things most of the time. This can explain lot of choices that happen.
the thing is, it doesn't suck for a lot of people. I definitely have some minor issues with it, but it encodes the workflow i was already using in GNOME 2 as the expected workflow.
Fine if you like it, but this is literally the first I've ever heard of anyone suggesting that GNOME 3 was a favorable continuation of GNOME 2; you hear that far more about MATE.
I'm hoping Chromebook will win the Linux Desktop wars. I'm not sure what it's future is, but right now with Wayland, running on GNU, running on Linux characterising it as yet another variant of KDE / Gnome / Xfce or whatever doesn't seem too outlandish. It runs Debian GUI packages just fine - for example Mozilla says to run Firefox on Chromebook you just install the Debian package. All the other Debian packages I've installed like vim and LibreOffice just worked too. It's pretty seemless as any Linux GUI app you install just appear in your startup folder just like native Chromebook apps do. It's fast, as in native fast, because it's an LXD container.
Google's control of the platform is almost Apple like. All the hardware has to work, the manufacturer has to supply 8 years of updates, and the battery has to last 10 hours. Compatibility issues vanish because in effect it's like buying a windows laptop or macbook. You're guaranteed the kernel can drive the hardware.
Security wise it's on par with a phone, which means it's better than any desktop OS (Windows, MacOS, and the traditional Linux), which for me is a major plus.
The downside is hardware availability. If you are a student who only needs 4G RAM and 64GB SSD, you're very well catered for. You can get something amazingly cheap that's build like a truck and has a waterproof keyboard. But finding something with 16GB or more of RAM and 256GB+ SSD is like searching for hens teeth. Here's hoping that will improve.
Clearly the open source funding model is broken and companies keep pushing experiments into projects, or trying to insert themselves into the ecosystem
Most OSS developers are not well compensated, and get frustrated
Also does not help that people are prone to jump into the hypes, like Wayland, clearly a worse experience for most users
Generated hundreds of thousands of bugs, for application developers, package/distro maintainers, and end users to try solve it. Pressure putting everyone on edge. I am sure many people will probably get PTSD after this
We don't need unbounded advancement. We need something that retains the core ideas of OSS and moral integrity of people involved . What we're seeing now is what it should look like. We don't need open source MSFT or Apple with their ever advancing progress into exploitation of all the low hanging fruit.
Clearly that guy seems unnecessarily confrontational
One Wayland OSS project leader attacking, and reporting, another... ouch
I mean OSS developers already get abused by companies spending hours doing free work, now they also have to spend hours managing community feelings, and drama between projects? Should be unnecessary
Your comment reminds of Haskell's motto, which may apply here as well.
"Avoid success at all costs" meaning, not to intentionally avoid success, but if you try to gain success/widespread adoption by compromising your other goals and principles, you have lost your way.
