Linux is super fragmented (and generally has been historically).
If more people in Linux agreed to develop, use, and support the same distro–similarly to how most of us use the same kernel–then that distro would probably be better than Windows and more people would move to Linux.
I get what you mean from a development perspective, but I already know your opinion is and will expectedly be unpopular within the Linux community. People here like freedom and choice, in one way or other.
A big benefit is that it keeps threat actors away from most components. If there’s a vulnerability in some Qt library, it will be big enough that people will notice eventually and fix it, but too small for someone to write ransomware code if the target is like 15000 random unimportant people.
There are certain things that people are trying to improve upon, to make things uniform/easy like the many ways to install software! Package managers to avoid having to compile from source. Flatpak/Snap/etc. to avoid having to deal with packages and dependencies! AppImages to put everything in one portable binary! However, with increasing simplicity you trade off customizability and other factors. This is why all these ways are available.
Linux isn’t as fragmented as you think.
Part of the beauty of Linux is the variance, though. There can be strength in diversity.
Yeah, we could also end war and hunger and change a lot of other things if we’d all agree to do it but here we are …
No no no no no no
It’s not fragmented in that sense.
No, it would not be more popular if there was one distro. That wouldn’t solve any problems.
This is a fundamental misunderstanding I see among many people, especially those outside/new to the Linux world. They talk about having too many choices where it’s overwhelming to pick from. But it’s not like ice cream flavors where “Oh I like chocolate and vanilla and strawberry? How can I choose?” where some people can get choice anxiety; if you think that’s comparable you just don’t know about how Linux fragmentation works.
Now, that’s fine; not everyone will know everything, and this concept is not always obvious to everyone. That said, an argument made from ignorance is not a valid argument.
Let me explain why fragmentation doesn’t work this way.
In every piece of software that is fragmented in the Linux world, it’s not arbitrary. It’s not people making hundreds of different things “just because.” There’s always a correct choice for each person. Different tools in the same area to suit different needs. No, not all tools are on equal footing lacking unity. They all benefit from the same standards but implement the features that matter to that tool. Unifying them solves nothing. We may not even get a tool out of it as people would fight over the directions of the projects.
For instance, why are there different DEs and WMs? Because not every DE has the workflow a person wants. I can’t stand the Windows way of UX; I think it’s terrible. If there was only one distro, and it came with KDE, I’d be very frustrated as there’s no good tiling options!
The different distros are not ice cream flavors; they exist to fulfill specific needs. You pick your distro, DE, etc to suit the way you want to use your computer. Everyone has a way they want their computer to work whether they realize it or not, even if that way is just how Windows does it.
There’s not an overwhelming amount of distros; that’s a view stemming from a lack of understanding. Fragmentation is not a problem.
then that distro would probably be better than Windows and more people would move to Linux
So as you can see, this wouldn’t be the case. That distro wouldn’t serve people’s needs, just like Windows doesn’t serve people’s needs.
The problem with Linux is not its fragmentation - that’s it’s superpower; there are distros that will meet the needs of everyone already. You just have to figure out what you want from a computer. If it’s just how Windows does things then, well, there are DEs and distros out there already made to function like Windows! Give Mint a try, for instance.
The reason Linux isn’t more popular has nothing to do with not having a good-enough distro that can beat Windows. What that looks like is different for different people, and I guarantee all of them exist somewhere.
Tbh Linux already is better than Windows (and Mac) on every front except two:
- Lack of industry-standard software for certain fields as well as a handful of specific games
- Normies will use whatever their PC comes with and will be too scared to reinstall, and Windows and Mac come on almost all devices by default.
P.S.
similarly to how most of us use the same kernel
This isn’t the case. We don’t even use the same kernels!
First, many distros use very different versions and second, some come with kernels that have major tweaks and customizations.
Not to mention the various modules and kernel parameters that get enabled and added.
There are plenty of kernel tweaks.
While I broadly agree, I feel like the focus of these fragments is just scratching their own personal itch, rather than building an improvement. I quickly lose interest in “software that only values X” vs “software that only values Y”. I just want software that’s good.
The problem with that is simply that what counts as good will vary from person to person.
I think there’s a baseline that Linux developers need to take more seriously. The PopOS people are good reference, especially their upcoming DE.
Imagine you had a basic desktop; top bar, dock and desktop widgets.
Each of those is a different app and each app comes in 4 different flavors: “Fast but ugly”, “Pretty”, “Tinkerer’s dream” and “Well designed but under developed”.
Sounds great right? Just pick whichever ones you like best. But along the way, things that would be considered “requirements so basic we don’t even need to state them” are not met.
So you value pretty and get the pretty versions. But each of them was developed by a different team, each with different opinions about what “pretty” is. Your desktop doesn’t have a cohesive look with colours and fonts mismatched in a way that no monolithic project would ever tolerate.
So you grab one pretty app, and two tinkerer’s dream apps. You can make them match yourself! That’s the power of Linux. You put a week of work into bringing your desktop up to this default standard, fighting a mountain of faff along the way. Each app uses a different language for their code and configs. Each app supports different features. You find yourself wishing they followed the “well designed but underdeveloped” app, but it hasn’t had an update for years.
Finally, your desktop is ready for unixporn… as long as you don’t open your file browser.
Well yeah, but that’s what standards are for. Look at Wayland. Outside of GNOME being a bit slow, all the major compositors and DEs like KDE and Hyprland have agreed to implement certain common desktop features that every desktop should have along with the Wayland protocol itself. Then they go their own way.
So it’s not really as you say. There is unity in development beneath the heavy diversity.
But along the way, things that would be considered “requirements so basic we don’t even need to state them” are not met.
Except they are
Your desktop doesn’t have a cohesive look with colours and fonts mismatched in a way that no monolithic project would ever tolerate.
This is a bad example bc it’s opposite to what you say. You haven’t set a universal theme. Theming is commonly supported across desktops to some degree. You can get all your apps to have the same look unless an application forces its own (which would happen even in a homogenous Linux world). Some desktops will do it for you, but if they don’t it’s still as simple as install a universal theme, apply the Gtk version, and apply the Qt version. My understanding is soon it will be even simpler once a few more Wayland standards get adopted.
We already have the best ecosystem. It literally could not improve functionally
I think you’re right, but I don’t think most people who use Linux care about whether or not it’s popular. Popularity is almost entirely irrelevant to the philosophy.
I disagree; I think that we do care about it being popular enough that it incentivizes software and hardware vendors to support it rather than ignoring it.
It would stop being linux.
Yes, but my choice of distro is the correct choice of distro everyone else needs to switch for unification reasons.
Yeah nah, your distro sucks ass, my distro is the best!
You both are wrong! I wrote my own distro with opinionated modifications and it is superior to both of your distros!
How dare you
Plot twist: you’re running the same distro
First off nobody can agree on anything, ever, so this is some fantasy world that will never exist.
Second, I’ve seen this complaint for at least 20 years and yet the Linux ecosystem is stronger than ever.
Fragmentation is a strength, not a weakness. There are amazing developers who would never have had the opportunity to contribute if development was monolithic like you are proposing.
Linux is already better than Windows, the latest versions are a mess, and is likely going to get worse.
As long as it does not have systemd and I can use either Wayland or X11. Easy, no?
What is the issue with systemd?
SystemD is fucking great and i will never go back to anything else, and this argument is the reason OP doesn’t understand why a unified agreement on a distro is fucking impossible.
To be fair, there’s already a general consensus in which are the “beginner friendly” distros, which includes the community support and familiarity with windows. Most people would probably recommend Mint to a new or switching user, for example.
After someone is used to a more basic linux distro, diversity and complexity are pros rather than cons. “Oh you need a distro with super specific specs for a niche use case? There’s 5, you can pick or try them all”. I could have linux on every machine i own, including a TV or even a thermostat, and I’d have a different setup.
Ha. nope.
I thought we all agreed on NixOS? ;)
No.
Amen brother
Nope.
Debian is The Universal Operating System.I think a lot more people are jumping ship than getting onboard as of late.
Jumping ship for what?
It’s the right way to build an OS, IMO.I’m totally committed to flakes as the right way to do dependencies after Docker got everyone doing it wrong.
Perhaps a fork like Aux, or to Guix
Sure they are, if your sources are drama articles calling the death of nix when 15 out of almost 4000 maintainers leave…
It is indeed an effective tool to scare people. Too bad it works
