The topic of downstream and upstream is an important one in the Linux ecosystem, where from one base distribution you can go many layers of distros deep before even looking at all the other base distributions. Within that veritable jungle you get questions about who is responsible for packaging software, where to report bugs found with a specific application, as well as what ‘LTS’ truly means in a consumer context. These and other points are raised in a recent video by [Brodie Robertson], with many examples of things going tragically wrong.
There’s a good argument to be made that ultimately it is the distro that is responsible for the software that they provide via their repositories. As [Brodie] shows in the video, there are a few cases where an ‘LTS’ distro uses an old version of some software that contains a bug that has been fixed a while ago, so reporting it to the developer is rather pointless, while the distro maintainers should fix it with backporting of patches or updating the version.
From an end user experience this also makes the most sense, as in the end they just want to have the Windows experience of downloading a proverbial installer, clicking through whatever dialogs pop and have working software. If the software is provided via the distro, it is their responsibility, the same way that you contact the developer if you get a DEB or RPM from a GitHub project page and it doesn’t work.
This current Linux Chaos Vortex can be called a major issue when e.g. FreeBSD has no such upstream/downstream issues, with cross-platform installers being basically impossible on Linux ever since the Linux Standard Base effort died.
Perhaps Linux will get a distroless future, however, which may finally herald that Year of the Linux Desktop.
LTS is more “we’ll fix some bad security issues for a little longer” it doesn’t mean that all bugs will get fixed, most will be left.
If your in a corporate environment, it’s easier to support as the bugs this year are the same bugs as last year, not new bugs every week.
The distribution is of course responsible for the distribution. If I wanted the latest App I’ll be going to the developers site and grab it there. If that does not work I might switch distributions.
I passed on Mint for the same reasons mentioned in the video. While I liked the idea of a nicer Ubuntu, Mint is no Ubuntu ;)
I don’t want to be switching distributions simply because a single app I want doesn’t work. What a hassle!
It kinda parallels the old song…
“Look again, it will say
All you do is “plug and play”
Why do I spend every day?
Reinstalling Windows
“
Or build from src
Look again, they will say
It just works out of the box
why do I spend every day
recompiling the kernel
It is a hassle. Mint was supposed to be the least amount of hassle. But for me that did not work out (I just deleted many sentences), leave it at that.
/rant on
I despise snap and the Ubuntu app store. It is morally wrong, technically wrong, architecturally wrong, poorly implemented, poorly maintained.
Now it adds another hurdle, that must be overcome, by whoever tries to provide a solution to what Ubuntu was supposed to solve.
It also shows a vanity that managed to inject itself into that foundation, that is not based on expertise, instead it just tries to lazily and superficially copy what corporations have done to make profit, absurd.
I’d consider Ubuntu the worst distro out there (non-rolling, outdated packages, breaks on updates, king of dead ends (snap, mir, upstart, …)). Distros typically have their own repositories, user repositories and the possibility to install 3rd party package managers (e.g. nix). Developers offer AppImages, Flatpaks or language-specific installers (python, rust, go, …). Usually there is a plethora of ways to install the same piece of software instead of the other way around.
“in the end they just want to have the Windows experience of downloading a proverbial installer, clicking through whatever dialogs pop and have working software”
Or use whatever centralized repo they have on their mobile.
It works because of the Android virtual machine approach.
Imagine if every phone model basically needed its own repository for compatibility. You’d get the situation of the old Symbian phones where you had to hunt high and low to find apps made for that specific version and variant.
This problem was identified more than 15 years ago. I can’t remember which developer called it out, but they basically made the argument that Linux has a fundamental “casting flaw” in the fact that each distribution is trying to own all the software on it as a middle-man between their developers and their users, and this is severely limiting the scalability of the platform into popular use.
Hi, there’s also that “binary application packaging” part at
https://en.wikipedia.org/wiki/Criticism_of_Linux#Desktop_use
Funny quote: ” in 2006 Dominic Humphries stated that the aims of the Linux community are not desktop market-share or popularity, but in Linux being the best operating system that can be made for the community.”
The irony is that providing for the needs of the wider community and not just the core developers or die-hard geeks would lead to Linux becoming very popular and gaining a large market share on the desktop. The fact that it doesn’t just means that the “Linux community” as identified by Humphries is not actually practicing what they preach, or he was defining it in deliberately narrow terms that are basically just “me and my buddies”.
The basis of your argument has a major flaw.
“The community” referred to isn’t ‘every human’. It was never intended to be and shouldn’t be.
Being the best for the community of users who are willing to learn something about the system they use, take responsibility for setting it up, maintaining it, and using it in a specific way, is a completely different thing than what you assumed the original argument is.
“I want to make the best microcontroller IDE and learning environment for students and makers” doesn’t mean it should cater to my local baker who wants a push-button solution.
It also doesn’t mean it should cater to someone who is already comfortable laying out boards and writing in assembly.
There are certainly awful neckbeards who lord knowledge over ‘noobs’ and damage the community.
But the bigger problem with Linux adoption by ‘normies’ is that it isn’t made for them.
The VAST majority want an appliance. The will actively avoid learning when they can. They want nothing to do with setup or configuration.
THAT is why phones have been, and will continue to be the primary computing device for people.
They work, unless they don’t.
Deal with it.
And when a new one comes out it gets replaced.
CAN you take apart a refrigerator and tweak stuff? Can you replace parts? Can you repair anything significant?
Sure! But it is an appliance so you don’t.
Box make cold.
No more cold? New box time.
“Popular use” pffff, so it will be a target to mainstream regulations? Anyone whos not dumb should probably switch to linux with the world getting too familiar with the cyberpspace and irl laws infecting it. Last thing I want in that era is anything i care about to be popular. Mainstream is the worst thing in existence.
Indeed, the push to standardize such aspects of the Linux experience with the LSB in the early 2000s was an attempt to fix that. Of course, by ~2007 that had bled to death and subsequently we saw an explosion of new distros, all trying to be its own island. We could have had literal ‘Linux applications’ like we have Windows and Mac and Haiku/BeOS applications, with a much clearer ownership and responsibility.
The distro repository model is a golden cage with locks and guards just the same as any appstore controlled by a big capitalist corporation. The difference is that because it’s maintained by the community that you choose to identify with, there’s an illusion that it’s maintained for your benefit and that there are people who are listening to your needs as a user.
The reality is the complete opposite, because the people who actually maintain the repository and all the software in it don’t have any interest in you. They don’t care if it doesn’t have the software you need or want – all you do is make demands and annoy the people who have more than enough work in maintaining the software packages they want to use.
I think you need to distinguish between distributions. Some are more corporate and stable, but no longer free. Some are community wild west and still free.
The same problem persists whether you’re on a “stable” distribution or not, because getting software through the respository demands developer effort to maintain it there, and that is a scarce commodity, and doesn’t scale beyond some thousands of packages or applications – and most of those are going to be just operating system components anyhow. The offering for end user applications is going to be slim pickings.
One of the issues is that it’s anti-competitive. You have competing offers for the same basic application where each is trying to best the other to sell more or solve some problem the other isn’t, but from the repository maintainer’s point of view that’s just duplicated work.
Related:
https://blogs.gnome.org/tbernard/2019/12/04/there-is-no-linux-platform-1/
https://blogs.gnome.org/tbernard/2020/03/25/there-is-no-linux-platform-2/
Don’t use ruzzian Linux distros, especially popular ones like CachyOS
Evidence of this being a Russian distro? While the founder and lead developer have Russian names, they’re based in Bavaria. Which is in Germany. Also, the CachyOS website is actively blocked in Russia. Odd for a “ruzzian Linux distro.” How about you put down the paranoia and do some actual investigation?
The youtuber does not understand security.
He seems to be a bit more of a user activist. Aggregates well.
Still trying to make this “youtube a day” huh
When your job is replaced by Laotian software engineers who (with support of LLM) will do your job for 10% of typical EU wages, you gotta find alternative ways to make ends meet. This includes writing slop for Hackaday.
Obligatory: https://xkcd.com/2347/
I will believe in AI when it can diagnose and fix software incompatibility issues. Also when it can find and install the a working distro and drivers for my laptop.
“The worst in the world … except for all the others.”
A basic principle of the Free Software Manifesto, often forgotten, was that software should be free but support should cost. If you don’t want to vet and build every package yourself, the solution is supposed to be to actually purchase that support, either by buying a productized version like Red Hat or by hiring someone to do that vetting for you.
RMS assumed that open source meant that you could get that production level support from many places, but that probably the best support would come from the authors since they at least theoretically understood the code best.
You may not always get what you pay for, but don’t assume you can get what you aren’t willing to pay for.
Before you enthuse too much about single distribution, remember just how complicated it sometimes was to set up the build parameters for each architecture, and sometimes each system. Config swept a lot of that under the carpet, but there was always the risk that your system wasn’t one that the provided configuration logic already anticipated and handled.
The best argument in favor of the Apple ecosystem used to be that it was standardized enough, both hardware and software, that everything just plugged in and worked and was supported. The disadvantage was that this was only true as long as you were satisfied with what was provided within that core ecosystem, and the assumptions broke down when you tried to go beyond that. For some people, that was and is a good trade-off. For others, less so.
Think about what sort of incentives that ethos creates:
Make software only so good that people will use it, but not without paid support. In other words, make half-assed software.
I liked the shareware concept. Users could get in touch with the author/programmer and pay a registration fee if they liked the software.
In return they got a printed manual, often. They also could ask for source code, sometimes it was already being offered in the registration form.
That way, they could support the author and encourage the author to improve the software.
There also was public domain (PD) software, freeware, postcard ware etc. These were simpler times.
Meh. I loved shareware as a kid although I rarely had any money to buy it. Now I could but what is available free in the OSS world is so much better than anything I ever saw from shareware….
Depends. In German, there’s an old saying that too many particpants can ruin something.
It’s “zuviele Köche verderben den Brei” wgich translates to “too many cooks ruin the pap”.
That’s one aspect of open source software that’s both negative and positive.
It also leads to all those forks (which are ultimatively beimgbabandoned), if people can’t work together and pull on same rope.
As much as I like the idea of team work and working together peacefully,
I also think that the individual should be valued.
And just seen as one of many nameless contributors.
In the past, in the days of shareware/freeware/public domain, you had very gifted indiviuals with great ideas that worked on something all alone.
It was their work, their brainchild.
They were open for input, but it was up to them to decide how things develop.
That was the time of the bedroom programmers in the 80s and the one-man-companies in the early 90.
When friends or family members (father, son or brothers) wrote software for DOS, Windows 3.x or Atari ST/Amiga.
By contrast, open source doesn’t value the original author or the project as much.
Anyone can take the source code and turn it into something else, can pervert it or use it for something that violates the original’s author’s intention.
An example for this are open source autopilot programs for Arduino or Pi hobby drones.
Such software, being open source, is now modified, used in warfare, used to harm people.
It’s used against the original intention of its author(s) which was to bring joy and happines to people.
But Linux/FOSS is blind to this. It relies on applications being available as source code, because closed binaries aren’t well supported.
There’s no official binary compatibilty among all the distros, no true backwards compatibility.
Open source software is no option on Linux, it’s a requirement, an enforcement, a religious principle.
Windows world is different. It has open source applications but also a standarized binary compatibility (PE EXE, Win32 API).
That’s why wonderful things like Kylix had no future (the Linux’s Delphi).
Binaries are shortlived on Linux.
Good software that relies on intellectual properties of others can’t be used.
Commercial DLLs made available by others can’t be used, as it’s possible on Windows.
Back in the 90s, third-party VBX DLLs supported in Visual Basic 3 powered a whole software generation.
Hobby programmers could easily integrated VBX files into their own application, which was very user-friendly and allowed rapid software development (RAD).
There was a whole market for it.
Developers supported developers by providing their VBX DLLs to others.
Either for free or for a small fee.
That way, developers could earn a bit of money and felt valued.
Their work supported their living, basically.
And unlike with open source software, they worked on their software. They didn’t have to become service staff in order earn money.
They remained programmers, did what they loved.
^And NOT just seen as one of many nameless contributors.
Shareware wasn’t “pay if you want”, it was gimped or time limited until you paid the fee to unlock it. Some software was fully functional, but kept nagging at you for not paying. Different methods of inconveniencing the users until they paid money were attempted.
The whole idea was that you didn’t spend any money putting it on the store shelves, because the users would distribute it for you. That was the “share” part.
No, that wasn’t the spirit of shareware! That was “crippleware” or demo software.
Some shareware authors did do that, yes, but real shareware had no limits.
Real shsreware had shown a polite notice by the author that asked you to please consider to register after a brief period of testing (30 days?)
Shareware with a nag screen (timer) also existed, but it wasn’t good practice.
Same thing, different name. The point of shareware was only about who was allowed to distribute it. “Trialware”, “Nagware”, etc. are sub-categories of shareware. The original term was “user-supported software”.
The original point was that prior to WWW and cheap hosting services, small independent software vendors didn’t have good access to distribution channels, so Shareware was born by allowing users to copy the software between themselves without accusation of piracy.
Point being, if it was fully functional software with no limits and no nags, it was freeware instead of shareware.
Sometimes people group “donationware”, under shareware, but it’s really the same thing as freeware. If there’s any sort of repeated reminder or time delay with a splash screen that tells you to donate, etc. then it becomes nagware again, which is a type of shareware.
Dude, I’ve been there in the 90s and I had many shareware disks and CDs.
I do assure you that shareware had rules, it was being mentioned in the shareware catalogs on those media. Go check, if you want.
And crippling software was NOT the idea of shareware.
Authors who did limit their “shareware” were not playing by the rules, they simply took advantage of the shareware distribution.
“Freeware” was often something different. No registration fee, no printed manual, no source code, no extras.
It was like Public Domain, except that the author didn’t give up on his author rights.
Yes. If the only motivation that existed for creating Open Source software (or hardware for that matter) was to sell support then it would all be very shitty indeed.
Here in the real world we have hobbyists, corporations and even governments contributing because they want the software to exist for their own use. In some cases that use is free hobbyist stuff, in others they are using it for their actual business. (Yes, I realize that the days of the majority of the Linux kernel being written by hobbyists in their mother’s basements have passed) Either way most are selling neither the software itself nor the support. They are doing it for entirely different reasons.
Just take a look at how KiCAD has progressed since CERN became involved. Is CERN selling KiCAD support now? I very much doubt it!
Incorrect. Cern started up a paid support program three or four years ago, in lieu of direct donations.
On the subject of KiCad, I have wasted many weekends attempting a succesful source build since version 7. Recently tried with 9 – still no go. FWIW, my dev box is Slackware.
Relying on hobbyism alone doesn’t make for a good software ecosystem either.
At the end of the day it comes down to paying people to do a proper job – not paying people for their hobbies. For the latter, you go to Patreon and open up a tip jar, so people can pay you to noodle around with some interesting project for years and years without really going anywhere with it. There the incentive is to drag your feet for as long as possible, to have an excuse to beg for more money.
That is partially an issue with writing software for sponsorship as well – when the project is done, no more sponsorship, so the smart software developer will always leave a bit more work to be done…
Meanwhile, when you write software to sell the software, the incentive is to get it done as quickly as possible, and better than your competitors, so you could then sell as much as possible.
There are of course caveats, like gaining a dominant market position or using other tricks to put down your competition without putting effort to the actual work, enshittification etc. but the general principle of it is better than “We won’t listen to your complaints because it’s already free.”
There is no doubt about this.
Just check your crap before you ship it, it’s a core principle in software development.
If it’s broken, fix it. if you can’t fix it yourself, take it down or leave a big red disclaimer saying “IT’S BORKED, WE’RE WORKING ON IT, USE AT YOUR OWN RISK”.
Yah, sure. For the easy to use, beginner distros that’s a pretty reasonable request.
I’ve been using Gentoo for 20 some years now. I like it a lot and don’t regret it. But it’s naturally going to be a bit more complicated than that and I don’t blame the maintainers.
First… many ebuilds ‘packages’ have all sorts of options the user can pick. I would expect that on an otherwise stock system a maintainer should be reasonably certain that none of these options break. But it has to coexist with and depend on scores of other packages. Those programs and libraries each also have their own options the user may have changed. It would be impossible to test every combination!
Then there is the mixing of testing and stable packages. I have never seen a distro that had everything I wanted with every feature I wanted all in stable packages. But I don’t want a fully testing system either! So there is always some mixing. That testing package should have worked on the maintainer’s system before they checked it in. But has the exact same mix of testing and stable packages as exist on my system ever been tested by anyone else?
Finally it’s a rolling release. Rolling releases are WONDERFUL! But not perfect.
My desktop was booting slow and throwing all sorts of error messages that went by so fast I couldn’t read them and didn’t seem to affect anything. But last night I finally went through them all. There were all sorts of things trying to load that had been deprecated years ago. That kind of stuff builds up when you never have to do a reinstall. It went from taking several minutes to start to less than 20 seconds!
So, anyway… if you want something where everything just works without tinkering.. you have to choose a distro that is meant for that. It’s great to have choices!
“they just want to have the Windows experience of downloading a proverbial installer, clicking through whatever dialogs pop and have working software.”
What is this guy talking about?
In my experience installing something from a distro’s package manager is FAR more likely to “just work” than downloading a random installer from a third party regardless if it to run in Windows or any other OS. And usually there are less dialogs popping up. If it’s a good distro there are none!
I think he just wants Linux to be Windows for him. But with penguins.
Yes, when to contact the package maintainer vs upstream is an important discussion to be had… for the power users. For the people who are installing bleeding edge testing systems or mixed stable+testing. What’s the Windows equivalent? Install a bunch of beta and/or alpha releases? Is that even a thing anymore? Or would you maybe have to grab latest from Git yourself and start building?
Convince me that installing Linux and/or building from source should be easy and accesible to the masses. Humanity generally sucks; most people are stupid (myself included). Let’s not make stuff worse.
Hopefully I will be worm food long before the ubuntu trash becomes the next windoze.
I will stay within my simple-minded insular bubble using Slackware until my wife decides she’s had enough of my stupid shit and arranges for my demise.
I think this article is fundamentally just about the tension between desire and reality — i.e., a mind game. Like, this whole concept of “responsibility” is over-used in almost every context, but especially in software, and even moreso in Linux…like, what? I mean, obviously someone who is paying for a support contract from RedHat will have a different perspective on this but i am using free and open source software. The only “responsibility” i want to assign to other people is that if i have their binary based on a GPL project, then i ought to be able to get source to their modifications. And i have to accept that that responsibility goes unfulfilled too.
I love linux because i have been able to get to the bottom of almost every problem i have had for 30 years. Before, with DOS or Windows or MacOS (before X), problems were insurmountable, or only amenable to guess-and-check techniques. This idea that there is such a thing as a Windows app that just works on every Windows machine is fantastically counterfactual. Even official updates from MS don’t work on every machine!
But if you don’t enjoy getting to the bottom of problems, what does Linux have to offer you? I’m sad to see people complaining about this situation. If you don’t want Linux’s superpowers then there’s not much point assigning blame for your own desires.
But i guess it’s really a consequence of the fact that there’s nothing else. People are using Linux these days because the things it’s weak at, it is sometimes better at than any competitor.
Then again, how many of those problems are simply caused by Linux being a poorly mess? How much of your “Linux superpowers” go to simply maintaining the operating system itself or getting poorly made software working?
That’s a nirvana argument. It’s meaningless to demand such absolutes. The fact of the matter is, software compatibility backwards and forwards on the Windows platform is light-years better to the point that people don’t really need to think about it. Meanwhile, Linux doesn’t even have a platform because it refuses to define and maintain one.
of course every problem i have is because of awful software. that’s a given. all software is awful.
in the brief moment in my life where i had to use windows, i encountered constant bugs and constant compatibility problems. and everything was young back then too, i can’t imagine compatibility has gotten better.
imagine a world where windows has half the bugs as linux but you can only fix 10% of them. you just imagined a counterfactual: windows is way buggier than linux. come on, man. i get being frustrated but let’s be honest about the status of mainstream software.
Though if you define “every Windows machine” as the NT family running on x86 hardware that isn’t broken and can successfully boot to the desktop, then yes there are Windows apps that run on every single one of them. Most likely you can include Windows 95/98/ME in there as well.
Here’s a guy pulling apps out of the Windows NT 3.1 ISO and running them on Windows 11:
https://www.youtube.com/watch?v=DKdI6HmILUA
They work just fine, because in Windows, an “app” doesn’t need to integrate itself into the operating system hierarchy like it was the Borg. An app at its simplest is just an executable file plus whatever DLLs it needs in the same folder, because the system is standard enough to do that.
Linux haters society meeting again today, eh? I hate to break it to you, but even Microsoft uses Linux for its servers. As do Amazon, Google, and all the other big players. Microsoft and Apple have far more problems with their insular systems, than Linux with its diversity. They constantly have to steal from open source, such as BSD for OS X and Open SSH used in Windows. Just like in biology and evolution, diversity is strength.
Yes, for their servers.
What does that have to do with desktop user software distribution?
With Linux there are many options you can pick. It’s freedom. Also in daily use, I’m finding Linux far more stable in every way than Windows machines I constantly have to keep fixing at work due to software and hardware issues.
They are slow, full of crapware starting from Microsoft Edge all the way down to OEM software and popups. Some data stealing software (“user experience improvement”) always running in the background too. 8 GB RAM on Windows? Barely usable system. Linux with 4 GB just runs.
Software distribution? APT from Debian/Ubuntu/Mint just works. It never failed me in 15 years.
Haters? Where? Are those users being meant who still have their own mind and haven’t been given up on critical thinking? :)
Critical thinking means rationaly checking choices.
One is free, stable, community driven and is environmentaly responsible (20 years old desktop, no problem).
Other is led by greedy companies, is not free, requires you to throw away perfectly fine 8 year old computer, and is becoming more and more problematic due to spyware and illegal data collection.
Even EU is waking up with requests for sovereign cloud. OS is also an obvious sovereign issue. Consider why is that the case.
Have a great day.
Generally speaking, extreme views aren’t helping, either, I think.
Neither are enforced religious beliefs or heroism/people culture (pple, Lnux)..
Open Souce=good, closed source=software bad is sort of black&white thinking.
There are good reasons for both principles, I think.
It’s like with capitalism vs socialism, I think. Neither is entirely good/bad.
Both have some ideas that are worth thinking about.
So chosing one in its entirety is extreme, while demonizing the other leads to extreme views.
Finding a healthy balance can be a goal, maybe.
It’s like with Yin and Yang, maybe. Things don’t always have to contradict.
Also importantly, if one system entirely “wins” and one “looses” , people have nothing to compare with anymore.
In worst case they then stop questing the status-quo and stop self-reflecting, stop searching for further alternatives, stop improving things.
That basically happened when the iron curtain fell in the 90s and capitalism had “won”.
It became a blue print for many places with out questioning it (anymore).
I’m glad that where I live we have a social market economy from before that time.
It’s not perfect by any means, though. It’s flawed, too, and we know that. We keep critizising it, too, which is good.
Same should be done with Open Source and L*nux, I think.
Never stop asking things. The more (constructive) criticism and self-reflection the better. It’s part of any truthfinding process.
And here’s the problem, these companies can contribute to the Linux kernal, can infiltrate it.
They can do that exactly because of open source.
Democracy is good, but if it can’t defend itself it might be hollowed.
There’s no easy way to fix this, maybe. Vigilance of the people might be required.
Not paranoia (no witch hunts), no, but keeping a bit of scepticism and raising an eyebrow once in a while.
Same goes with L*nux and Open Source.
Supporting for what it stands for is surely good,
but so is to never trusting it entirely. Just because it has good label.
(To use an analogy, more than often, the bad guys won’t introduce themselve as bad guys but rather as good guys.)
Or let’s put it this way..
Open Source is great as long it remains a free choice among others.
The day it becomes mandatory, though, is worrying.
Because once it becomes enforced, developers working on free software might be required to use their real names, too.
A total, fully enforced anti-obfuscation act in the future might require that.
In a democracy, that in turn is like forcing voters to openly state for which party they do vote for.
Their anonymity is taken away from them in the name of freedom and transparency.
That’s how 1984 type stories do begin.