When will MS replace the NT-kernel in windows?

[JohanH]

Isn't it looking at what is actually hitting the metal?

At this point the "Unix philosophy" referred to has become a "Unix religion", because it's not based on technical facts or merits, rather on dubious political nonsense. All I see in that camp is bitter old farts, somehow making technical development cross their dusted political views. Just spare me the BS.

[nctnico]

p.s.: Two years ago MS hired Lennart Poettering (systemd inventor) who was working for Red Hat...

Windows is going to be even more fucked than it is now.

I don't understand the hatred for this guy. And I've followed the whole systemd debacle since the start, read his blogs etc. I guess, he was too clever and too outspoken for the mob. It for sure brought out all the ugliness you can imagine in the circles; the amount of trolling and hatred by his critics I haven't seen since (except in politics). But what it takes to have development and invention (and society) make huge jumps is guys like him that are visionaries and able to realize the ideas, too. I for my part admire him. If anything, he will do good things at Microsoft.

I guess if he should have respected the general "Unix" philosophy that tools should do only one thing and do it good,
the reception should have been much better. Instead, he created a monster that wants to do and control everything.
What will be his next step? Creating a Linux registry for all settings like windows does?

That wouldn't be bad perse. From a maintenance point of view, it is easier to have a singular way of interacting with a system compared to needing to edit several files in several places using several different formats. Look at the sendmail configuration as an example (although I think this program has been created by aliens given the way it needs to be configured).

[shapirus]

I don't understand the hatred for this guy.

It's very simple. In the context of systemd/journald he: a) went on fixing things that weren't broken; b) did this in a way absolutely hostile to unix philosophy. Either of these is as much of an antipattern as it can possibly get. This, naturally, resulted in creation of a hard to use and troubleshoot abomination, which, worst of all, destroyed any possibility of using an alternative solution in all mainstream linux distros.

Hope he now finally feels at home in microsoft. Pity that he did not leave redhat before damage was done.

p.s. pulseaudio, OTOH, was actually a useful thing, one that actually made the world better.

[shapirus]

That wouldn't be bad perse. From a maintenance point of view, it is easier to have a singular way of interacting with a system compared to needing to edit several files in several places using several different formats.

Unless you want to automate things and set up monitoring, or you are simply an experienced user who understands how to take advantage of the so-called unix way. Plain text logs and configs, separate configs where each file is responsible for a small part of the whole, the "one small program for each task" approach are all better suited for automation and for power users.

Windows-style binary logs, configs, and monstrous software systems that try to control everything are natural and welcome in the windows world. They like them there, so keep them there. Don't bring this BS to where it's not welcome, it's as easy as that.

[magic]

Now the future is clear.

Lennart will fork Linux and Windows will become the first operating system based on systemd-kernel.
Nobody will be using Ubuntu by 2030, mark my words.

[JohanH]

It's very simple. In the context of systemd/journald he: a) went on fixing things that weren't broken; b) did this in a way absolutely hostile to unix philosophy.

I strongly disagree. People are entitled to their opinion, but referring to the "Unix Philosophy" doesn't bring any technological arguments. It's not even an agreed upon standard, but "philosophical" arguments. But talking about beating a dead horse. As for a), he was completely free to fix whatever things he wanted. Anyone can write any software that they want. Most people agreed it was great and here we are today. If it's not good anymore in a few years, maybe something else comes along and replaces it. Who cares, it's just a bunch of 0s and 1s. But the "philosophers" had to turn frigging software into some kind of religious debate and hating people for it. 

[Karel]

Lennart will fork Linux and Windows will become the first operating system based on systemd-kernel.

 

[shapirus]

I strongly disagree. People are entitled to their opinion, but referring to the "Unix Philosophy" doesn't bring any technological arguments.

I don't agree. It's not just someone's preference, but a set of practical considerations which create specific advantages, some of which I have already mentioned: easier setup, automation, troubleshooting, reduction of requirements for interactive management. There is a disadvantage too: you must know how to cook it right, otherwise you will end up with a tangled mess instead of a powerful tool (sendmail mentioned in previous posts is a good example).

There is nothing religious in this. It's purely technical reasoning backed by experience.

...otherwise, I agree. Everyone is free to write whatever they want. The problem however is when distro maintainers begin forcing their preferred solutions on everybody without offering an alternative. Well, those who want can create their own systemd-free distros. That's how Devuan was born. BTW I actually use it on an embedded system where systemd was undebuggably slow or freezing on shutdown and had some issues on boot. Too bad it doesn't have as many users as mainstream distros, meaning less maintainers and a smaller community.

[PlainName]

It's very simple. In the context of systemd/journald he: a) went on fixing things that weren't broken; b) did this in a way absolutely hostile to unix philosophy.

I strongly disagree. People are entitled to their opinion

Except when they don't match yours, apparently.

but referring to the "Unix Philosophy" doesn't bring any technological arguments. It's not even an agreed upon standard, but "philosophical" arguments.

Just like programming patterns, no? 'Best practices' similar.
Besides which, shapirus has given you excellent technical reasons (which you surely already knew but chose to ignore, because).

[shapirus]

The problem however is when distro maintainers begin forcing their preferred solutions on everybody without offering an alternative.

...speaking of which: when will MS replace the tall icon-based crappy Windows taskbar with a no-nonsense button-based taskbar where each button represents exactly one window and have a well-defined action on click: raise a minimized or inactive window or minimize an active window?

Before that happens, in case anyone is looking for a solution, like myself, who was speechless (actually no: I did say words that I won't repeat here) after seeing the windows 11 UI for the first time and finding out that it wasn't possible to disable it, here's what I can recommend:

- proper taskbar: https://github.com/dremin/RetroBar
- proper start menu: https://github.com/Open-Shell/Open-Shell-Menu

I spent the effort of searching for them, installing and setting them up even though I use Windows very rarely in a virtual machine. That's how extremely bad and unusable its newest UI is.

[JohanH]

Except when they don't match yours, apparently.

It's the constant bashing and seemingly hatred of Lennart Poettering that I don't like, not people's technical opinions. When you are unable to keep it at the technical arguments, it becomes a poisonous and disgusting environment.

[Karel]

Regarding the list of facts presented in the starting post, we can add:

- MS hires Lennart Poettering, the inventor and lead programmer of systemd.

- MS released Edge webbrowser for Linux

Also interesting:

In October 2018, Microsoft released 60,000 patents to the Open Invention Network members for Linux systems,
but exFAT patents were not initially included at the time. There was, however, discussion within Microsoft over whether
Microsoft should allow exFAT in Linux devices, which eventually resulted in Microsoft publishing the official
specification for open usage and releasing the exFAT patents to the OIN in August 2019.

and:

Paragon Software Group opensourses their (until then closed/proprietary) NTFS driver for Linux without any objection/protest from MS.

Technical difficulties aside, all this justifies the question when MS will replace the NT kernel with a Linux kernel.

edit: added Edge webbrowser

[shapirus]

Technical difficulties aside, all this justifies the question when MS will replace the NT kernel with a Linux kernel.

It will be Lennux, a proper Linux rewritten from scratch.

[Bud]

Efking hell no, computer boot time will be back to nonsense.

[asmi]

Technical difficulties aside, all this justifies the question when MS will replace the NT kernel with a Linux kernel.

Not anytime soon for sure. Linux kernel is a total mess compared to modern NT kernel, and their propensity to break binary compatibility on just about every kernel release turns driver maintenance into a nightmare, while for NT kernel, even quite old drivers tend to work just fine. Couple that with incredibly slow and convoluted process of upstreaming a driver (which is the only semi-reliable way to ensure your driver won't break down after every other kernel release), and you can see why Windows model is much more appealing. One can get a device driver from zero to release in a couple of months, good luck achieving the same for Linux driver - you'll be lucky if somebody bothers to even look at your PR a year after submission, and then will complain about some BS like a semicolon being in a wrong place, and you will have to go through entire process again While Windows kernel interface is mature and stable, and so you can simply compile your kernel once, get it signed - and you're done.

[Karel]

So, it's not true that, when windows Vista was introduced, new drivers had to be written...

Binary compatibility inside the kernel is only important if you support closed source drivers which Linux doesn't
so it's a non-problem. Except maybe for nvidia graphics cards but I'm convinced that if MS would switch to a Linux
kernel, also nvidia would switch to open source drivers like AMD and Intel already do.

[nctnico]

Technical difficulties aside, all this justifies the question when MS will replace the NT kernel with a Linux kernel.

Not anytime soon for sure. Linux kernel is a total mess compared to modern NT kernel, and their propensity to break binary compatibility on just about every kernel release turns driver maintenance into a nightmare, while for NT kernel, even quite old drivers tend to work just fine. Couple that with incredibly slow and convoluted process of upstreaming a driver (which is the only semi-reliable way to ensure your driver won't break down after every other kernel release), and you can see why Windows model is much more appealing. One can get a device driver from zero to release in a couple of months, good luck achieving the same for Linux driver - you'll be lucky if somebody bothers to even look at your PR a year after submission, and then will complain about some BS like a semicolon being in a wrong place, and you will have to go through entire process again While Windows kernel interface is mature and stable, and so you can simply compile your kernel once, get it signed - and you're done.

That is why sensible manufacturers of hardware  (or software needing special drivers) distribute their drivers as source code which gets compiled during installation of the supporting software. Take Virtualbox and Nvidia as an example. All the infrastructure to do this exists from the kernel side.

[shapirus]

Not anytime soon for sure. Linux kernel is a total mess compared to modern NT kernel, and their propensity to break binary compatibility on just about every kernel release turns driver maintenance into a nightmare, while for NT kernel, even quite old drivers tend to work just fine. Couple that with incredibly slow and convoluted process of upstreaming a driver (which is the only semi-reliable way to ensure your driver won't break down after every other kernel release), and you can see why Windows model is much more appealing.

Ironically, Windows NT kernel is much more of Unix way example than Linux, which is indeed an absolutely monstrous monolith now.

You're absolutely right about drivers, and that's a big problem in real life, which must be obvious for anyone who has not isolated himself, like e.g. RMS, in a bubble where nothing but FOSS exists and nothing but FOSS has any importance. In reality, the practical impossibility of distributing binary drivers that will work across a wide range of Linux kernel versions is, in addition to userspace problems of a similar nature, a significant hindrance to its wider adoption on desktops.

[Zero999]

What's the point in binary drivers? The vendor makes money from the hardware, not drivers and they can still distribute the driver source, under a proprietary licence if they want to.

[langwadt]

What's the point in binary drivers? The vendor makes money from the hardware, not drivers and they can still distribute the driver source, under a proprietary licence if they want to.

They might have bought some of the code from someone else so they can't opensource it. The are are afraid the source code would reveal some of their secret sauce in the hardware

[nctnico]

What's the point in binary drivers? The vendor makes money from the hardware, not drivers and they can still distribute the driver source, under a proprietary licence if they want to.

NDAs and so on. But having binary drivers is not much of a problem. It is easy to provide a bit of 'glue' software that binds a closed source driver into the Linux kernel using a loadable module. There is a crapload of drivers which do this. Almost every graphics driver is a binary blob for starters.

[Nominal Animal]

It's the constant bashing and seemingly hatred of Lennart Poettering that I don't like, not people's technical opinions. When you are unable to keep it at the technical arguments, it becomes a poisonous and disgusting environment.

I don't hate Lennart Poettering.  I'm sure if I met him in person, I'd be happy to have a coffee and talk shop about him.  I hope he has a good life.
I do hope he stops all software development work, and never does anything involving Linux again, because he does more harm than good.

I would like for the large companies (Red Hat, Microsoft) and distros and his lackeys –– those who believe that because something is new and created by a socially adept woke person like Lennart, it must be better than the old stuff by those 'racist misogynist long-bearded smellies' –– stop using social and business reasons to push his crap onto all Linux users, and actually look at the miserable quality of his code.

You obviously haven't looked at it yourself, nor the history, nor the CVE reports involving his code.
Even PulseAudio was horrible crap at the beginning, refusing to co-operate with anything else, until others stepped in and fixed it.

Lennart Poettering is very adept at manipulating non-technical people, but his development skills are rather lacking.  He also has a very high tendency to reinvent a different shaped wheel –– often defective ––, and throw crap at people who complain.

The quality of his attempts at modify the Linux kernel was so bad Linus told subsystem maintainers to not take anything from him (until the quality gets significantly better).  This started Lennart's hate towards the Linux kernel developer community, and lead to very difficult situations and fights.  For example, at one time he decided to hijack the kernel command line 'debug' term (so that when set, systemd would emit so much boottime debug messages kernels with default dmesg size would fail to boot), arguing that 'the kernel does not own the kernel command line'.

His 'kdbus' design idea –– moving the dbus into the kernel –– was so horrible, that even though it was submitted by Greg Kroah-Hartmann (a very friendly, very good Linux kernel developer; done a lot of work to help manufacturers and vendors upstream their drivers and changes into the vanilla kernel), it was deemed to be utterly horrible crap.

Do your own research, and do not accept your social feelings as assumptions as the truth here.  Do not dismiss my statements here just because you dislike me.  Start at say mitre CVE list involving systemd, and look at the mailing lists during the development and adoption of these: it is not good.

In that light, the vitriol is warranted, in my opinion.

The Unix philosophy in this context refers to the tested-and-true design philosophy centered around minimalism, modularity, composability, simplicity, and separation of mechanism and policy, instead of monolithic opaque all-encompassing frameworks like systemd.  Monolithic frameworks are insular, difficult to adapt to different use cases, and suffer from single point of failure; and their security and bug track records are worse than those of smaller components, because of the much higher internal complexity.

You can think of it however you like, but the fact is, it is one of the rare design principles/philosophies that has weathered the test of time: it works well.

There is a reason the Linux kernel and the GNU userspace components that make up for the core operating system, can be used in anything from tiny appliances to supercomputers (all 500 most powerful supercomputers run Linux currently).  Modularity and configurability is at the heart of that.  (Even though the Linux kernel at run time is monolithic, it is highly configurable and consists of selectable hierarchies of subsystems and modules itself.)
Monolithic single points of failure like systemd mean for example that where in the early 00's I could upgrade my system without rebooting, I now have to reboot regularly like a miserly Windows user after a patch Tuesday.  I don't like it, and it messes with my options and workflow.  That, too, in my opinion warrants at least some vitriol.  Targeting it at the product instead of the instigator would be completely silly, like blaming water for the hole in the dam.

[newbrain]

The last thing I need today is some religious debate, so I'll just leave this here.
https://youtu.be/o_AIw9bGogo?si=9ce2LL3SUVHKOZRU

[Nominal Animal]

The last thing I need today is some religious debate, so I'll just leave this here.

The last thing I need today is a passive-agressive asshole who instead of considering the argument, paints it as irrational religious debate, and attempts to elevate themselves above it by linking to a video on youtube where someone else says their opinion.

You, newbrain, should be ashamed of yourself.

[tooki]

Regarding the list of facts presented in the starting post, we can add:

….
Technical difficulties aside, all this justifies the question when MS will replace the NT kernel with a Linux kernel

No it doesn’t. Those things are orthogonal.
If anything, many would argue that Microsoft is acting in its own self-interest by promoting compatibility with its technologies in competing products. That compatibility means Microsoft products stand a higher chance of being selected than before, because everything can work with them.