A multi billion dollar company with massive investments in Linux making an almost workable solution means everyone else is screwed
Microsoft expects Windows to be a means to an end: You run Windows to use your computer.
Linux neckbeards expect Linux to be the end to a means: You use your computer to run Linux.
If Linux is fragmented so bad that it is completely incompatible with the way software development and support work in the real world, the problem is Linux because Windows, Mac/iOS, and Android (incidentally a flavor of Linux) can all deal with it.
Of course, if you're not interested at all in mainstream desktop Linux adoption and are content hacking away at FOSS code while grumbling about the evils of capitalism and proprietary code, then more power to you.
Nowadays, with cloud platforms making the underlying OS transparent, even less.
Are you seriously telling me these billion dollar corporations can't manage to get in contact with some popular distribution's maintainers and work something out?
Nope, because those Linux distros aren't making them any worthwhile money while sending over far too many worthless end-user complaints.
If Linux neckbeards truly want to realize the Year of the Linux Desktop, they have to accept how the rest of the world at large works and play by those rules. It's how Android obtained mainstream success despite being Linux, and it's something any other Linux distro can do if they ditched the neckbeard pride.
Or to put it another way: The vast majority of computer users don't care about free-as-in-freedom or open-as-in-auditable source code. The only thing users care about is getting shit done. All other operating systems, including Android, understand and respect this. It's only Linux that chooses to be either willingly naive or in denial.
They can do it, but it’s not guaranteed/unlikely to lead to mainstream success. Android had a multi billion dollar corporation behind it, and likely needed that to succeed.
Another Linux distro that freezes its ABI and supports commercial software distribution likely would be irrelevant until it got hundreds of millions of users, and getting there is a big challenge.
Firstly, name-calling does not help in getting your point across.
The opposite argument could easily be made. What would be the point of the "year of the Linux desktop" if Linux is not substantially different from other OSes in the way it treats its users? That's why nobody is celebrating the "era of the Linux palmtop" with Android.
Linux makes different trade-offs from those made by the commercial OSes. The diversity is valuable. That's not to say there isn't room for improvement, but I would be pretty bummed if Linux lost what makes it different.
The point I want to convey is not so much that Linux should change (though as a desktop user I certainly wouldn't mind), but that anyone within Linux who complains about How The World Is Wrong(tm) needs to wake up from their freedom-infused obsession and smell reality. The rest of the desktop world functions fine, so if it's only Linux that Just Can't(tm) then the problem is Linux.
Linux is working fine for those who want to use it as such. No neckbeard owes a 'fuckstain' rest of the world a god damn thing.
Any access to Linux subsystems or syscalls on Android, work as matter of luck on specific devices, or is an area Google isn't yet enforcing Android security userspace rules.
Linux has suited me very well as a desktop for 10+ years. (I'm aware it is probably not suitable for every person or use case.)
To maintain the health of the Linux desktop(s), we do need to be open to new people and ideas from outside. But Linux is, and should remain, different from proprietary OSes. Otherwise, what is the point?
It's like saying a certain screwdriver must be made with a smooth ball point to be different from other screwdrivers. Nevermind that nobody can figure out a use for such a screwdriver and everyone happily (or begrudgingly) goes back to using flathead, Philips, and Torx drivers.
Most people don't care about free-as-in-freedom software or open-as-in-auditable source code, they just want to run Office and Photoshop and maybe play some snazzy games. A tool must first and foremost be useful in order to achieve mainstream appeal, Linux has consistently failed to do so because it is flat out not useful for most people.
If achieving the "Year of the Linux Desktop" is a real goal of the Linux community at large, some fundamental changes in ideology must happen:
* Acceptance or at least tolerance of proprietary source code (eg: Nvidia drivers). Most users don't care what philosophy of code they're running, they just want their computer to work and be useful.
* More emphasis on GUIs and a refined user experience. Neckbeards might only want the CLI and consider anything else below their ability to care, but most users want a good GUI like in any other widely accepted operating system.
* Accept that there can be such a thing as too much choice. Developers don't want to look after their code on five dozen flavors of Linux, some consolidation and stabilization of distros and runtime environments are a hard requirement to beating the chicken-or-egg problem.
* Less hostility to new users and outsiders. The elitism within the Linux community at large is stupid. Being a CLI wizard doesn't give anyone a higher horse to ride on, and it's not going to attract new users anyway.
* Less marketing emphasis on liberal and FOSS ideologies. They're all fine concepts to have, but most users care about free as in free beer, not free as in freedom. Cater to what users want in order to win users over.
Linux can be many great things, but its ideologies and philosophies hold it back in a wider world that values usability and practicality more than freedoms and openness.
If "Year of the Linux Desktop" is not a goal, then Linux can continue as it currently is. The enterprise world will be Linux's stronghold for the foreseeable future as before, and the desktop world will continue moving on with Windows/Mac/iOS/Android like always for better or worse. But Linux can't then complain the desktop world keeps disregarding them, because priorities are different and Linux chose to be incompatible.
I'm just one linux user. But my impression is that the "the year of the Linux desktop" is a joke even in the Linux community. I don't really care about world domination for Linux (although that's almost what we have on the server side).
I want a healthy enough market share for software availability to be good. For me, it is pretty good. I had to spin up a Windows VM for the first time in about 5 years last weekend to set up my doorbell with the manufacturer's utility. Otherwise there are more than enough games that run well on Linux to occupy the time I have (which isn't a lot). All the basics are covered and everything else is on the web. Dev tooling is second to none.
I want a useful, powerful, general purpose computing environment where I can do what I want without unnecessary obstacles, whether or not that what I want is aligned with the business model of a vendor. The philosophy is abstract, but in my view it has a real impact.
No software is "free as in beer" to create. I really don't know, but I wonder if marketing it as such would be a bad idea. Free software projects often need contributers as much as users. Companies with bigger marketing budgets want their stuff to be valued.
Otherwise, I agree with most of what you said about onboarding new users and less technical people, as well as fragmentation and ABI stability. Some of it is a necessary trade off - if people have the freedom to make changes to their software, you're going to end up with more diversity compared to ecosystems where that's not allowed.
Nvidea have announced the open-sourcing of their driver for Linux compatibility. Perhaps the GPL die-hards were right to hold firm in that...
The term "neckbeard" is an unkind stereotype. Please reconsider your use of it.
I have no plans on reconsidering calling certain parts of the Linux community neckbeards. If they want to force their ideologies down other peoples' throats ("you fix your intellectual property problems", "proprietary code is evil", "use the terminal", etc.), I'll call them how I see fit. Respect is earned, as they say; and they certainly aren't respecting the users nor the world at large.
That aside, objectively it can't be denied that Linux doesn't satisfy most users' needs. If it did we wouldn't be having this discussion nor would Windows have 80~90% of the desktop market. You're fortunate to be someone that desktop Linux can sufficiently provide for, but not everyone is like that.
Windows/Mac/iOS/Android for all their faults can satisfy the needs and desires of almost everyone, it's something Linux critically fails and needs to improve on regardless of worldly aspirations.
Nvidia open sourcing large portions of their driver was a welcome turn of events, though like you I'm not sure if credit for it should be given to the anti-proprietary diehards... :V
Software companies: OK then we won't ship our software for Linux
linux users: (surprised pikachu)
like, what exactly is the "excuse" here? FSF set a deliberately onerous license on the assumption that you will either join (because you need to use the GPL codebase) or do it yourself... and companies either do it themselves, or don't release software for the platform. Or they do and it breaks.
Same story as with ZFS basically. The license that is necessary for commercial games/software to run in the necessary ways (that require interaction with the kernel) is incompatible with the license that FSF has chosen. And you can't build an anticheat without poking at the kernel, otherwise it's trivial to hide from it.
Anti-cheat is fundamentally a problem of controlling the code that can run on an end-user's system - looking at memory or packets to scrape out useful data that the game is not presenting to the user - and that's functionally incompatible with a free software system in the first place. And attempting to do so requires interacting with the kernel, and if you're not in the kernel tree then you're chasing the kernel ABI. And anti-cheat rootkits will never be in the kernel tree anyway, period.
Where is the excuse? These are just incompatible products at every level, both conceptually and legally/licensing. That's by design, that's what GPL is intended to do as a copyleft license.
People seem to have this weird idealistic view of GPL, that it's about "protecting user freedoms" and that just makes all the problems of the world go away and everybody happily falls into line, but the mechanism by which it works is by causing problems and removing interoperability for developers of third-party software with incompatible licenses. If you don't do GPL, you can't play with the GPL codebase, and if your kernel is GPL and you need to do kernel things, then as they say - "wow, sucks to be you". But that's working as intended: GPL is a license which is intended to cause breakage and hinder developer freedoms, strategically, in the interests of greater end-user freedom in the long term.
If you just want open-source, the linux kernel should have been MIT/BSD licensed and it wouldn't have been a problem. But GPL isn't about open-source, it's about pushing back on copyright licensing as an ideology.
Which is to say:
freebsd users: wave
The ZFS situation here is a completely different (as is past bad behaviour by Nvidia/VMware) kettle of fish and doesn't need to be dragged into via a license discussion on an ABI thread.
The kernel keeps an evolving ABI as expected. This is a win pro/con for Linux, but frankly its so rare to hit an incompatability at the kernel layer it's as good as stable. More things break elsewhere typically many times before this is an issue on MacOS or Win1x for that matter...
As for anticheat what's that even got to do with Linux' API situation is beyond me.