[Raion]

I never was one of those radical people who was anti proprietary software. Yet even today I'm starting to find myself even more faithless when it comes to proprietary software. My main desktop operating system for the last 6 months has 99.9% been Windows 10 LTSC. 

I've mentioned before all the things that I don't really like about modern Linux, and these things go much deeper than systemd. The operating system overall has been pushed towards object modeling, mandatory access controls, complex firewall solutions, and more and more abstraction away from the nuts and bolts of things. Some of the most common systems that I cite are:

Udev
Dbus
*Kit (polkit, consolekit, etc)
GRUB2
Modern glibc

And more. As I've kind of stated in the past I don't have anything against dynamic device filesystems, or standardized IPC or any of the other ideas that these have come to embody. I don't like how they have implemented them. 

IPC on UNIX really only was standardized and handled well with STREAMS. It's compliant with the UNIX philosophy, it's not complex, and it's extremely well documented.

As of FreeBSD 12, devd is an elegant and simple device filesystem maintainer which can add and remove device nodes easily. 

Consolekit, policykit are trying to solve security issues that I think are being approached from the wrong angle. Consolekit sets device permissions and states for a multiuser login, but that's the thing, this was never a serious issue. Policykit also does something similar with permissions and privilege escalation. If you're going to talk about adding privileges to a user, the IRIX way worked fine. If you're going to talk about setting all these permissions it's probably better to approach it from the way that everybody else did for decades which was to add you or users to the system groups. 

Meanwhile, glibc has done a good job of trying to lock out many apps built against it from being used on other OSes or with other libc. I think that there are some cases where you can't even get complete desktop working without major patching for musl or uclibc... What's up with that shit? 

GRUB2 is just bloated and might as well be a mini operating system in and of itself

And now even the FreeBSD and other projects are falling victim to linuxisms and have for a while. Been watching the OpenZFS debacle closely; and while my expectations have been exceeded that's far I'm not entirely sure that the long-term sustainability of relying on Linux is going to go over well because as soon as changes that only work on Linux get blocked because of BSD and apparently, eventually illumos, it's going to probably cause a big fight where everybody ends up forking their code again and we will be back to square one.

GNU meanwhile despite increased friction with the Linux and rest of the FOSS community at large, and increased irrelevance of their killer apps, has continued to dig themselves deep into a hole. Not only does their code continue to get more and more bloated on mini applications, but projects like Guix and a few others are claiming to move towards Hurd, which 20+ years later has remained a singular architecture 32 bit operating system without any hope of the future. They are defending Stallman despite his defense of a well known pedophile, and one of their only allies, the free software foundation looks about ready to cancel them at the drop of a hat 

If the free open source software community can't make heads or tails of how to even begin to make the lessons of the past work properly and stop falling into stupid traps I don't have a lot of hope that there is even any point to being part of the free open source software community long term. 

Not only has GNU failed us when it came to being truly cross platform, as they have one of the most messed-up planned obsolescence schemes for removing support for operating systems, but their overall attitude has been toxic to anybody that doesn't meet their narrow definition of free software (i.e. GPL, no blobs, etc.) And that has weakened them in the face of competition.

[soviet]

Yes indeed, i ended up the same way, using windows.
Always have my hopes in HAIKU but also ended up stall in nothingness like hurd.

[lunatic]

sometimes it is good to just remember the history of Unix. Beyond the problems you describe - however one might think about exact details - the story seems to repeat itself.
There has always been an insane multitude of different systems, approaches, and ideas. Most of those do not prevail. Even though Unix started out with a simple structure (one tool - one task) it has grown larger and more complex than anyone really wants. This is not simply a GNU/Linux problem. You can see this across the history of Unix. There has never been a real vision of what Unix should be or how it should develop. Instead, there has been an over-reliance on individual people, some of whom - although admittedly brilliant - have been lacking in terms of social competences. This has sometimes led to something like a reverse-justification: When inadequate documentation is justified by saying that Unix should be for professional users only and if you are unable to understand this, you are just too stupid to use Unix. Or when developers violently fight over design choices, but nobody actually cares about the majority of users. Unix development has always been like modern social media. Whenever something is hot and new, everybody jumps onboard and it grows and gets a lot of attention. Then at some point, people lose interest for whatever reason and 5 seconds later, you are an idiot when you are still using it. Sometimes the importance and influence of ideas can be judged only looking back. Take 4.3BSD for example. It's a highly important and hugely influential piece of software that basically nobody used - or did you have a VAX? But pieces of it were adapted, developed, redeveloped, and found their way into a lot of later systems. Additionally, as you mention free software, this is not just about the software itself, but also about the idea of freedom. Certainly "freedom" is just a word and what people mean and understand when they use this word can vary greatly (bigly? ).

A very old joke goes like this:

The different versions of the UN*X brand operating system are numbered in a logical sequence:
5, 6, 7, 2, 2.9, 3, 4.0, III, 4.1, V, 4.2, V.2, and 4.3.

Sometimes, you just have to accept the internal logic of Unix and choose your poison.
Me? I have been a happy/unhappy Mac user since MacOS X 10.2. It works most of the time in most cases and I can always have some terminals open too.

[Raion]

Sometimes, you just have to accept the internal logic of Unix and choose your poison.
Me? I have been a happy/unhappy Mac user since MacOS X 10.2. It works most of the time in most cases and I can always have some terminals open too.

The biggest issue is that this idea you have a choice is an illusion. BSD, illumos, Solaris, etc. are increasingly dependent on GNU/Linux derived pieces, that often work subpar. And it's not even a necessarily massive issue with the source project, rather it's laziness. I've talked with a few people, and it's possible to build Firefox and GTK3 without dbus with some code patches. Why do people not do it? And why do I hate GTK3?

People don't do it because they're either lazy or one of those that just doesn't care what the underlying components are. The latter is fair, but that projection affects everyone downstream in a negative manner.

I hate GTK3 because it's gone in a far more Windows-like direction with how it runs, and removing dbus basically causes multitudes of programs that use atk calls to crash randomly. That's the biggest issue.

As for macOS, it's not really a UNIX proper due to its strong divergence, and while I used it when I worked from home out of job necessity, it's not a good OS for a variety of reasons, at least in my opinion:

1. I don't care about the pretty colors. UI "prettiness" is secondary to functionality. Transparency, 3D effects etc? It doesn't matter to me. Genie effects? It all is ugly, and I usually turned the eyecandy down as far as it allowed me.

2. That segways into my next point. The OS holds you back from doing things. The same issue I have with mandatory access controls. "Oh, you can't do that!" idiot proofing isn't able to be turned off on macOS. Yes, it's useful for retards who may delete core files or change core settings irreversibly, but for me, the constant having to turn off macOS's "untrusted application" nagware/lockdown every update, the "oh you can't delete this file, it's necessary for Safari!" (I deleted Safari, or used to, on Tiger and Leopard because fuck Safari). etc.

3. Mach is a bad kernel. Performance, messaging, etc. are all poorly designed. Most of macOS isn't FOSS. CoreAudio and other core components of macOS are as poorly designed as Windows.

4. The hardware lockdown. Yes, you can hackintosh, but nobody should have to do that! That's ridiculous.

5. macOS's cultural apple-isms. Status symbols, you're immediately lesser of a "techie" in many circles if you use a nonApple phone or PC. They will circlejerk against you for hours about how their machine cost $2500, but probably has less than $800 of component and assembly costs. The whole culture is bad and toxic.

I get you and many others like it, but I'd almost prefer to use Windows over it; just because the overall situation is not any better, and in some ways, worse.

[commodorejohn]

Unfortunately, that's human nature for you. People hijack other people's efforts either because they want something without having to do the work themselves or because they hold some conviction that their way is The Right Way, and other people go along with it because they don't notice, don't care, or want to get in on the Movement for their own reasons, and once they amass a certain amount of momentum (be it social, technical, or financial) everybody who doesn't like it (even the originators of the work being hijacked!) can just go jump in a hole. And the FOSS movement is particularly vulnerable to this because the whole point is that there's no real way to stop anybody (even bad actors or misguided zealots) from doing it.

I'd still take any flavor of freenix (well, any non-systemd flavor) over Windows 10, though, and I say that as someone who was an XP diehard for fourteen years. That shit is a dumpster fire even now, to say nothing of the horrors back in 2016.

[micrex22]

Unfortunately it all comes back down to "developers developers developers". I have proprietary software that hasn't yet been ported outside of the 'big two' (Windows/MacOS), and there are a lot of bugs in the *buntus and open source software that doesn't seem to ever get fixed or deemed as "a feature working as intended" at some point. That's why I much prefer proprietary UNIX as the experience is more cohesive. I think Linux in the terminal and TMUX is very much a fantastic experience though--assuming you can get your software to run in the terminal.

I've stopped trying to have a "one OS to rule them all" and instead run: Windows 95/98SE/XP/7/10 LTSC along with OS/2, AIX, Linux and whatever else. Having a Linux laptop as a bed top is fairly nice though, runs light and can steam things without having to ramp up the fan constantly versus Windows.

[jan-jaap]

Not only has GNU failed us when it came to being truly cross platform, as they have one of the most messed-up planned obsolescence schemes for removing support for operating systems, but their overall attitude has been toxic to anybody that doesn't meet their narrow definition of free software (i.e. GPL, no blobs, etc.) And that has weakened them in the face of competition.

I think you need to read the GNU Manifesto

GNU was never meant to be cross-platform. GNU was about providing a single platform which would obsolete all commercial alternatives. A free (as in speech and beer) OS in a time when AT&T was increasingly closing up UNIX. If you want to create your own OS, you need your own toolchain, and, lacking that OS, that toolchain has to run on some other platform(s) in the meantime.

The GNU toolchain turned out to be pretty good, at a time when commercial UNIX providers were starting to charge money for development tools. The GNU kernel (Hurd) never went anywhere, but there's no such thing as a commitment from the FSF to support obsolete OSes indefinitely. Quite the opposite, after all, GNU was supposed to replace commercial OSes.

If it wasn't for the work of Rainer Orth, IRIX support in GCC would have died much earlier. This guy was almost single-handedly doing the maintenance and platform support for IRIX, OSF and some obsolete Solaris versions. Time consuming volunteer work, I may add. When he stopped and nobody stepped up, it ended. It's that simple.

[lunatic]

I think you need to read the GNU Manifesto

GNU was never meant to be cross-platform. GNU was about providing a single platform which would obsolete all commercial alternatives. A free (as in speech and beer) OS in a time when AT&T was increasingly closing up UNIX. If you want to create your own OS, you need your own toolchain, and, lacking that OS, that toolchain has to run on some other platform(s) in the meantime.

The GNU toolchain turned out to be pretty good, at a time when commercial UNIX providers were starting to charge money for development tools. The GNU kernel (Hurd) never went anywhere, but there's no such thing as a commitment from the FSF to support obsolete OSes indefinitely. Quite the opposite, after all, GNU was supposed to replace commercial OSes.

If it wasn't for the work of Rainer Orth, IRIX support in GCC would have died much earlier. This guy was almost single-handedly doing the maintenance and platform support for IRIX, OSF and some obsolete Solaris versions. Time consuming volunteer work, I may add. When he stopped and nobody stepped up, it ended. It's that simple.

I would second that almost completely, but add something:
The open source movement always had an ideological dimension too. One of the key differences between the GNU GPL and the classic BSD license is what one can do with the code. Whereas GNU was much more about "free as in speech", BSD was effectively "free as in beer". In this regard, the GPL had a certain appeal to developers who wanted to create a new, free ecosystem that could not be absorbed by commercial products and companies. As you said, the ultimate goal has never been to support say gcc on IRIX or Solaris, but to build up a new GNU operating system that is unix-like, but free. A lot of volunteer work was needed and people need motivation to do it. The GNU idea apparently provided this motivation.

[Raion]

I think both of you have misunderstood my point.

GNU drew resources away from BSD and other efforts for decades because it was "good enough" and because a generation of developers became enthralled by the GPL's anti-proprietary software views. Here's the thing: BSD licenses have to be respected too. You can't just take a BSD-licensed program and relicense it. That's illegal and against the code of the license. You have to retain and offer those files under the BSD license unless the entire file is rewritten, or it's incorporated into a proprietary project without source code.

The difference is that weaker licenses are willing to coexist with proprietary software, and accept, pragmatically so, that "it's gonna happen either way" as millions of Chinese, American and Japanese companies have violated the spirit of GPL for decades anyways, and Sony and Netflix donate thousands or even millions to the BSD projects they take code from.

Anyways, the point I was trying to make was more along the lines of nothing to do with IRIX or Solaris or anything in particular, it was a generalist statement that the way they choose to obsolete things is insidiously arbitrary, especially for proprietary software. It's all about how long the company chooses to support it, not if there's willing maintainers. AFAICT, over the years people protested the removal of IRIX from GCC and offered to step up and maintain it, but GNU said no. But that's one specific instance.

Back to the focus, by depriving alternatives of resources, we lost our ability to have true choice. People are beginning to realize this, which is why LLVM became a thing, and several other projects on smaller scales have stepped up to the plate. But it's not enough on its own. More must be done. If your non-GNU OS requires any GNU tools to be bootstrapped, that's a vulnerability. It's /okay/ if GNU can be used, by all means, but having them as the ONLY tool to be able to perform a function is a dumb and increasingly dangerous prospect.

While I'm against the philosophy, beliefs and views of GNU, I wouldn't be so vocal and angry about it if people had worked to maintain alternatives. That's the biggest problem I face, because while I can hack around code a little, it should be blatantly obvious from my numerous posts that I'm not the type of person who can design a compiler or a linker or anything spontaneously, let alone add support for an ABI or OS's eccentricities. I'd like to get there one day, but alas, I'm a bit dumb in that department. I hope this doesn't come off as entitled, because that isn't the point I was trying to push at all.

The idea is diversity. Choice. The ability to choose. The two don't have to be absolutely congruent or compatible, no no no, but they should be somewhat at parity. If your only other compiler lacks crucial features, that's an oversight, for example. That's why if I ever get to the point where I'm developing stuff, well, I'm going to if I have the money fund at least one competitor if the niche is only filled by me, even if that project chooses GPL. Why? Because there should be a diversity of opinions and projects.

Hope that clarifies things. Thanks for chiming in fellas.

[jan-jaap]

I think the sole reason Linux had any chance to take off at all was the shaky legal situation of BSD vs AT&T based UNIX in the late eighties / early nineties. Linux was an absolute joke in those days. But by the time BSD had rewritten all tainted (AT&T) code, Linux was no longer a toy. By then the WWW had been invented and suddenly everybody wanted to host a site without paying megabucks to SUN. These people didn't really care about GPL vs BSD license, but the idea that your server OS might be declared illegal will make most companies shy away.

Then again, I don't think you should have any illusions. If BSD had come out the winner, there would now be 150 BSD forks on DistroWatch, and many of the things you despise in Linux would have been invented in BSD. The thing is, OSes are far too complex to be developed and maintained by a couple of hobbyists. They need large teams of skilled developers and these need to be paid. In other words: you need a company to back it up, and that company needs revenue. It gets that revenue from other companies willing to pay, not from hobbyists and freeloaders. This means that they will develop features demanded by corporate customers, not individuals. You guessed it, these include mandatory access controls and complex firewall solutions. Because the treat from APTs (persistent threats, not Debian packages) is real. A breach of your servers can mean the end of your company. If your business involves military, avionics or space for example you have to deal with increasingly strict government rules and regulations if you want to participate in projects.