I have seen more people in the Linux community throw arguments that are weaker than flat earth "science" when trying to argue "distribution doesn't matter", especially in Linux. Each time I try to listen to these arguments I just get annoyed and annoyed. I hear the same things every single time. So I thought I would write another article, trying to finally end the "Linux Distribution Doesn't Matter" argument, one that is idiotic and quite frankly is invalid.
The people who promote the idea of distribution doesn't matter always and consistently fail to prove their claims. They always say something, explain a little bit, but fail to recognize gaps of logic. Let's debunk a few.
This one makes me the least angry, and I borderline agree. Keyword, borderline. You see, while the argument seems fairly solid, you soon get into a couple of issues. Like oh, I don't know... documentation, build tools, dependency support, licenses, bugs, kernel versions, kernel edits, and about a billion other things. While many people will try to argue this is irrelevant well tell me how.
As many might know, I work on Ubuntu Lumina, Lumina having no solid setup for Debian yet (except the okay-ish packages I made). I previously tried to work on Ubuntu Pantheon, until it finally made me give up because of how poorly documented it was and how terrible it was to compile. Yes, someone will put in the work to release it to everyone, but it is a lot of work. Also not to mention, some libraries might be absent or combined with other packages in other distributions. Distributions that modify their kernel will also have plausible issues depending on what they edit. Build tools might be broken or out of date for the package, so you will have to revert versions to older ones. See, this is not even a comprehensive list of issues people will have to face, and might not be willing to fight.
Also yes, many distributions will modify their kernels, also will have different utility suites. Some might have all GNU tools available, some with only the pre-included ones plus maybe some compilers in the repos, others will 100% remove GNU all together. Plus GUIs have their own pain with libraries, many using GTK or Qt, some using FLTK, and others making their own for who knows why. Speaking of GUIs.
The main argument here is that you can find any desktop environments in the repos of any distribution or source code to compile from. Remember how I mentioned Pantheon earlier. It isn't well supported for the Debian/Ubuntu tree, even though elementaryOS, the developers behind Pantheon is based on Ubuntu LTS. Not to mention that you might also not even find the desktop's source code well documented yet supported elsewhere. Oh and let's say there is theming, sometimes that is easier said than done to get the same themes, or config. This doesn't even mention window managers which may or may not work well, i.e. Openbox.
This one still hold drops of truth, I guess, but still insane of a claim.
So apparently there are people in Linux who fail to grasp there are humans who use Linux and hate Flatpak. Some people reply to Flatpak haters with dumb comments on how it is good or that we should improve it. How can we improve dumpster fires? Heck, I love Snapcraft, but many hate it. I even will not include Snap packages in Ubuntu Lumina by default (as much as I can avoid them) because I understand people dislike them, and they don't belong everywhere. Let's go down the list:
- RPM is awful and is not great to use, by all honesty. At least that can be improved, but I do not care to include RPM.
- DEB is great to install packages, but the packaging makes me want to cry.
- Pacman is fine if you want always updating, but isn't good at all for stability, and having an ISO update every week can make distribution developers want to stop. Look at Antergos.
- Others might not have plentiful sources for packages, and sometimes getting other distribution repos isn't as good of an option as many try to propose.
While there are distributions that aren't too special compared to others, some are often great. Let's look at Devuan, a wonderful systemd-less distribution. Often called a terrible idea, for no real reason. The main thing I hear against Devuan is that "it's a protest distro" and "what if Debian dies, what's the plan then?" and... anyone who says Devuan is bad or dumb is idiotic, as, like any Debian based system, there is a plan-B, or at least there should be. For instance, when Slackware wasn't as well produced for stable ISOs, SUSE (and in turn OpenSUSE) have continuation plans. So does Ubuntu, so does CentOS if somehow RHEL dies.
Back to the point, every Linux distribution is special in its own way. Even if it is plain Ubuntu just with satanic imagery added, it still is special.
And some people prefer starting places with more of an out-of-the-box workflow than trying to install everything themselves. This makes no sense to me as an argument. This isn't an argument.
What? I am sorry, this is false and idiotic. Who in their right mind, with any sense of economics and a basic understanding of the world would ever argue this?
This argument is so painful to me because it is just dumb. Yes, there are some research papers that show more choice is bad, but there is still a huge thing in the way here. Three main scenarios.
Scenario 1 - People do care about what they have. While choices are high and stressful, people will end up still picking something or building with LFS. In the end, still picking.
Scenario 2 - People don't care what they pick and will just jump to Ubuntu, maybe paying mind to other distributions later.
Scenario 3 - Distribution doesn't matter is correct, and we might as well stop working with Linux, as well as Windows 10 works fine enough for most people.
Yes, more choices might turn people off, but Darwinism is still at play too. Ubuntu Lumina 100% plans to promote itself through ads, and the main desktop OS I want to build (that is not a remix) will do the same. We should help each other but remember, those who do not adapt will die.
I have lost half the brain cells I once had trying to argue with these arguments. To the point, this article suffers from it. Yet, even if my brain cannot brain, my dumb ramblings are still saner than the argument that distribution doesn't matter.