I've cut out some parts because I have stuff to do. I haven't edited out things I wanted to ignore...
Go make an experiment. Take an average new laptop, try to install Ubuntu.
Have been running Debian on all of my machine including laptops for the past 15 years or so.. The last 5 years or so worth of laptops just worked after debian-installer had finished and I installed the non-free wifi firmware.
I don't think I've ever seen a Windows machine installed from the vanilla install media and not a restore disc that didn't need a lot of work to make all of the hardware work.
However, in the end, it doesn't change a thing. Linux doesn't work well for the end user for such reasons.
Can we get some empirical evidence of that and not just your feels on the matter?
Hardly ever for Linux. Guess why that is?
In the x86 world Intel, NVidia etc actively support Linux.. In the ARM world western vendors like TI, Atmel etc all have guys trying to get their drivers into the mainline kernel. You're a lot more likely to plug some random junk into a Linux machine and have it work than any other system IMHO.
The Linux problem is the "kernel of the day problem". Lack of stable interfaces within the kernel. Impossible to write a kernel module that survives a kernel release without requiring recompilation,
The "kernel of the day" that is probably the most widely used kernel out there if you ignore some deeply embedded RTOS like FreeRTOS and uTRON.
Kernel modules from one kernel version aren't meant to loaded into a different version?! Oh noes what a massive problem!
restructuring or fixup of its interfaces towards the ever-changing kernel-internal interface layer.
Which layer are you talking about? It's easy to make some vague statements but they are pretty useless here if you consider how many wildly different subsystems Linux is comprised off. One of the things I hear all the time is that the people that maintain stuff like the networking stack are too strict about changes i.e. refusing to allow hardware NAT support in. I'd really like to know which subsystem you think is constantly changing under developers feet.. maybe you mean something fairly recent like device tree for ARM?
That's a lack of a high-level statement "these are the interfaces, don't touch them for the next two years". It does not happen.
Linus has a high level statement "Don't break userland". As a user you don't care if the way the business logic in the kernel works changes as long as the exposed interface is the same or has glue to present the legacy interface to userland.
I hope you realise there are -stable branches of the kernel that you can use if you don't like actually seeing progress.. Considering you're fighting over an OS that's been dead for decades maybe those aren't old enough.
That doesn't make Linux irrelevant, of course. There is the server market
If you think Linux is just about servers you're either short sighted or being willfully ignorant.
If we start the same thing with AmigaOs, we again end up with the "Kickstart of the day" problem
AmigaOS is A: totally dead B: totally worthless C: Not interesting to anyone except for nostalgia. There is never going to be an agreement on what to do with it because there isn't really anything to do with it except tinkering.
The best thing that could be done is let the source code go and let people that want to mess around with it do so. I suspect the reason that didn't happen already is a little bit of "We need to squeeze every last drop out of this" and "We're not actually sure what we own, we'll make a big noise about things like the source leak but in reality if it did go to court we'd have a hard time proving what we actually own". If I was one of the groups that think they own AmigaOS I would be happy that someone leaked the 3.1 code as the torrents for it will probably out last all of the commercial entities that are trying to make a buck off of it and is less likely to be lost forever.
There need to be *some* agreement where the journey should go, and someone who makes this decision. I simply want to stop all this compatibility mess
Trying to make AmigaOS and Amiga hardware not a massive kludge of 80's and 90's crap barely hanging on for life in the present is never going to happen. If anything it goes against the spirit of Amiga. It was always a kludge and that made it lovable and interesting.
making everything for free and inviting the hacker of the day to create even more noise by creating another patch of a system component.
I would argue that if the source for AmigaOS and commonly used stuff like P96 was all up on github there would be less of a problem with people random hacking binaries and sticking them in unofficial patch sets. People are lazy and although there might be forks they would most likely be pretty close to each other. What you seem to want is someone to dictate who gets to release stuff, who gets to pick version numbers etc. That could happen if there was an official AmigaOS opensource release with an official maintainer, an official P96 opensource release with an official maintainer. I doubt there are enough people with the knowledge and time to manage that stuff left but it would certainly be a better situation than the only options being to disassemble existing binaries and release unofficial patches/beg the Amiga gods of long long ago for the right to fix their precious.
