Lennart Poettering goes to Microsoft

1. Microsoft is an ENEMY OF THE FREE SOFTWARE movement and I love the FreeSoftware movement.
Microsoft used to be a competitor of open source. It did what all competitors do: try to overtake and damage the competition. It did exactly the same thing to Digital Research, IBM, Borland, Sun, Oracle, Google and Apple.

Today Microsoft is a huge consumer of open source; it is probably the company with the 3rd largest number of Linux computers in the world (after Amazon and Google). It has understood that there is more value in using and leveraging open source than in trying to outrun or squash it.

Who is our friend? Do we have any friend between companies? Well, one sure i know is Google, it gives a LOT of resources to many FreeSoftware projects not directly tied to its immediate consumption.

Google is an even larger consumer of open source. Again, using the simple count of Linux instances, it has more than Microsoft (and that's really saying something!). But if you look at total funding going into Open Source (most of which goes into Linux, and the ecosystem around it), Google is not particularly large. Intel and IBM contribute more.

None of these companies are friends of open source. They are users of it, and in some cases (such as Intel) they have a symbiotic relationship with it. They know how to make smart investments.

1.3 Why Microsoft is enemy of the FreeSoftware movement? That is pretty easy, what pushes/pushed the majority to Linux or BSD ? That fact that Windows strinks!
The vast majority of all Linux machines on the planet (probably about 99%) are in data centers, being used as servers. They are there because their users (who typically spend B$ per year on computers) look at total cost of ownership; they don't care whether Windows stinks or not: if it were cost-effective, they would use it.

Ignoring servers: The use of Linux (and other free OSes, which are much smaller) as personal desktop/laptop machine is (a) much smaller than their usage as servers, and (b) much smaller than the usage of Windows on personal machines. You have to remember: Windows and MacOS together have about 90% market share on desktop/laptop machines! Even ChromeOS (an operating system without an operating system, as most people really just run Chrome on it) has a market share about as high as Linux!

So if Windows stinks, please explain why it is still the undisputed #1 in personal OS usage, with the Mac as #2, and everybody else far behind. Answer: It doesn't stink. To about 75% of all people who make the decision which OS to buy, a Microsoft product is the best answer.

3. Microsoft is a de facto MONOPOLIST, as such, ...
It is true that Microsoft has a near monopoly on the personal desktop/laptop OS market. But it is by far not the only monopolist in the computer industry. Linux has a near monopoly on the server OS market, and a 100% monopoly on the supercomputer OS market (look at the TOP500 list: every single of the 500 largest supercomputers in the world runs Linux). Google has a near monopoly in the search market. The market for commercial databases is not exactly a monopoly, more of a duopoly: between Oracle and IBM, it is pretty sowed up. If you look at Cloud, Amazon doesn't have a monopoly, but it is so much larger than any competitor (about twice Microsoft, 3x Google, and nearly 10x than any other competitor) that one really can't describe that market as a "level playing field". I can go on and on: If you want an enterprise tape drive, the heads will be manufactured by IBM, and the drive and library either by StorageTek (now a division of Oracle) or IBM. In disk drives, there are about 2.2 vendors left. The computer market is full of monopolies and near-monopolies. There are many reasons for that, too hard to explain here.

I like the way Esther Dyson explains it (about 25 years ago, when Microsoft's position was far stronger than it is today): Microsoft is not a dictator, it was democratically elected by computer users. The only problem is that the constitution of the computer world only allows for a single election. In reality, what happened in the intervening years: the Internet, and the movement of computers from personally owned single-user machines to fractionated servers in the cloud has really disrupted Microsoft; what is left to it is a near-monopoly on the desktop OS. In all other fields, it is a competitor, but so are many other companies.
 
… Microsoft Windows … I like to avoid it where reasonable. …

The same, with regard to avoidance.

It's essential to my work, I typically use Remmina on FreeBSD for RDP to Windows 10.

It feels like I make very frequent use of apps in this environment.

The reality is quite different (I know this mainly because one of the apps times out after something like an hour of disuse): use of Windows for essentials is not particularly frequent. This particular app is being phased out, then I'll spend even less time with Windows (more time with web-based Microsoft Intune) … and so on.

Web-based systems are quite a godsend.

Same for tablets wiping out the market for laptops for the past ~10 years. Any day now.

Via <https://noc.social/@AAKL/111913624271631140>:

Apple Vision Pro Review: $3,499 Headset Will Eventually Replace the iPad - Bloomberg

I doubt it.



to remove or change, can't see the button to trash the message, sorry

You can Report your own content, request deletion.
 
To do my actual work, I am a happy Windows 7 user.
I do a huge amount of work in commercial Photoshop (CS6) and Excel (2013).

I can't do Turbo Tax, Quickbooks or nVidia CUDA work on Win7, so I built Win10 machines and VMs with the crap stripped out of it.
I don't like the nanny in Win10, but have mostly worked around that problem.
I do not game.

I also do a lot of movie ripping with CUDA on a big Xeon, 128gb ECC, SSD and physical disks.
These are stored on a XigmaNAS Xeon machine running FBSD 12 and ZFS Z2
FBSD works perfection for my needs.
 
So if Windows stinks, please explain why it is still the undisputed #1 in personal OS usage, with the Mac as #2, and everybody else far behind. Answer: It doesn't stink. To about 75% of all people who make the decision which OS to buy, a Microsoft product is the best answer.

I believe there are three main reasons why Microsoft has owned the personal computer market since the 1980s. I'm sure you know already this, but you did ask the question.

( 1.) They made binding deals in the 1970s and '80s, with Intel, IBM, Radio Shack, and many other PC hardware manufacturers, to make sure that Xenix (an early Unix clone) and their "quick and dirty" (D)iskette-based (O)perating (S)ystems (DOSes), like IBM's (P)ersonal (C)omputer DOS (PC-DOS), (M)icro(S)oft's own eponymous DOS (MS-DOS), and Tandy's Radio Shack's "Tandy DOS," were REQUIRED to be installed, on ALL Intel based (P)ersonal (C)omputer (PC) microcomputers, and most of their many hardware clones. Binding, LEGAL type deals, contracts, & agreements.

( 2.) For most people, it's not much fun to INSTALL a computer operating system. A lot of PC hobbyists, and some of us professionals, might actually ENJOY installing OSes, but for most of the normal people I know, who don't work in "data processing" and/or "(I)nformation (T)echnology" (IT) businesses, or have a hobbyist's interest in building them, it's a painstaking (P)ain (I)n (T)he (A)nus (PITA) to load and install all those slow-loading diskettes, CDs, network-transmitted, or USB flash drive provided software installer programs. For such people, having a pre-installed OS in a ready-to-go piece of hardware is a distinctive plus, and gives commercial software vendors like Microsoft and Apple a HUGE advantage over FOSS systems which only rarely get to make pre-installation deals.

( 3.) Microsoft made binding deals with many or most PC hardware manufacturers to get the LEAD TIME they wanted to develop driver programs for ALL the innovative new hardware which was to be designed and sold in the personal computer market. Apple on the other hand had proprietary CONTROL over their OWN hardware. The competitive advantages of these strategies were clear, and pretty much INSURMOUNTABLE, although the PC clone manufacturers did manage to chisel into the hardware market for a little while, and some of them came up with their own, briefly competitive DOS clones as well. Microsoft still won out and its stock quickly made Bill Gates and his business friends among the richest billionaires in the world.

I've always believed these were the main reasons, and still do. They have a knack for squeezing out the competition.

And most of us here know that it's not always that easy to find FreeBSD-compatible hardware. It's the kind of challenge that sadly sends a lot of interested people back to Windows or Apple. Most people just don't have the time, or they've already purchased the hardware without realizing that it won't work with FreeBSD until somebody develops a driver. Then they realize how likely it is that nobody will EVER develop a FreeBSD compatible driver for their hardware. There's just too much hardware out there already, and newer hardware is being developed all the time.

I don't hate Microsoft, never have. Sure, I've tossed a few barbs at them now and then, for my own amusement, and to entertain friends, but I also managed to cajole a bit of a living out of working on their stuff, as well as on IBM's, DEC's, SCO's, and the FOSS community's stuff. In about 1985 or thereabouts, I had the privilege of typing a few commands into one of the earliest PDP-11 based UNIX systems which just happened to be installed at (S)outhern (I)llinois (U)niversity at (C)arbondale (SIU-C) at that time, and was told that most of Microsoft's software "innovations" were actually gleaned, second hand, by their early software designers having free access to UNIX development papers which were available to anybody who had the incentive to just walk into the school library at (U)niversity of (S)outhern (C)alifornia at Berkeley (USC-Berkeley), or one of the many other school libraries, at the many other universities where UNIX was developed by university students in collaboration with AT&T employees. So yeah, I do feel justified in calling them pirates, and all their defenders can take comfort in knowing that those pirates aren't likely to lose TOO much sleep over anything *I* have to say about them, what with their big fat bank accounts and all.

About that smell? Well, I do think Microsoft stinks a bit, MS-DOS certainly had a bit of a bad smell to it, and, well, Windows certainly works better now than it did years ago, in my opinion, but I still like the smell of (F)ree and (O)pen (S)ource (S)oftware a lot better.
 
Operating System market share 2018-2023:
Windows36.4435.8735.023.6029.7029.35
Mac OS X5.806.548.016.775.968.05
Android39.6039.0838.8340.7142.7939.63
iOS13.1113.9415.1816.5117.4216.96
Linux0.760.790.810.941.011.31
Windows is the clear leader in desktop operating systems. If mobile is included, Android and iOS, have a larger market share than desktops. Linux, FreeBSD, and others as desktop operating systems are niche and will never be anything other than that.

Surprisingly, mobile dropped share in 2023, but it will likely continue increasing. Excluding mobile, Windows and Mac have more than 90% market share.

Windows just works. People just want to sit down, plug in some device, etc. The majority are not technical, they don't want to use KDE Plasma and spend three years configuring it to churn butter and slice cheese. They just want to power on the machine, do what they need to, and be done. Windows does that for them.
 
Operating System market share 2018-2023:

Windows is the clear leader in desktop operating systems. If mobile is included, Android and iOS, have a larger market share than desktops. Linux, FreeBSD, and others as desktop operating systems are niche and will never be anything other than that.

Surprisingly, mobile dropped share in 2023, but it will likely continue increasing.
Looks like Microsoft had quite a bad year in 2021 if I read this right. Where did you get these numbers?
 
Systemd gained a new command named run0. Its purpose is to replace sudo with PolicyKit. Many people are not impressed, telling that "While it might be a good idea, since it comes from Poettering I won't even give it a try." To quote Poettering:

"There's a new tool in systemd, called "run0". Or actually, it's not a new tool, it's actually the long existing tool "systemd-run", but when invoked under the "run0" name (via a symlink) it behaves a lot like a sudo clone. But with one key difference: it's *not* in fact SUID. Instead it just asks the service manager to invoke a command or shell under the target user's UID. It allocates a new PTY for that, and then shovels data back and forth from the originating TTY and this PTY. Or in other words: the target command is invoked in an isolated exec context, freshly forked off PID 1, without inheriting any context from the client (well, admittedly, we *do* propagate $TERM, but that's an explicit exception, i.e. allowlist rather than denylist).

One could say, "run0" is closer to behaviour of "ssh" than to "sudo", in many ways.
...
The tool is also a lot more fun to use than sudo. For example, by default it will tint your terminal background in a reddish tone while you are operating with elevated privileges. That is supposed to act as a friendly reminder that you haven't given up the privileges yet, and marks the output of all commands that ran with privileges appropriately"

 
Systemd gained a new command named run0. Its purpose is to replace sudo with PolicyKit. Many people are not impressed, telling that "While it might be a good idea, since it comes from Poettering I won't even give it a try." To quote Poettering:

"There's a new tool in systemd, called "run0". Or actually, it's not a new tool, it's actually the long existing tool "systemd-run", but when invoked under the "run0" name (via a symlink) it behaves a lot like a sudo clone. But with one key difference: it's *not* in fact SUID. Instead it just asks the service manager to invoke a command or shell under the target user's UID. It allocates a new PTY for that, and then shovels data back and forth from the originating TTY and this PTY. Or in other words: the target command is invoked in an isolated exec context, freshly forked off PID 1, without inheriting any context from the client (well, admittedly, we *do* propagate $TERM, but that's an explicit exception, i.e. allowlist rather than denylist).

One could say, "run0" is closer to behaviour of "ssh" than to "sudo", in many ways.

Am I the only one who doesn't want a full env reset when doing single commands as root? I want my tools in $PATH, my shell init config, tmux env vars and whatnot.

Poettering's replacement is going in the opposite direction of what I want.
 
Well, as Poettering says, it will be more fun, you get a pretty color. Isn't that the most important aspect?

A long time ago, I read some post or email or whatever, saying that he knew a lot about C but nothing about system adminstration. On the other hand, most distributions seem to have embraced systemd, whether because of RedHat's backing or not, I don't know.
 

Attachments

  • 1714738770764.png
    1714738770764.png
    139.6 KB · Views: 33
Sadly, with recent naive ideas such as the new Graphical Installer, the actual target audience for FreeBSD is kind of diminishing further. It will alienate those who want technical simplicity and the Steam DRM platform gaming kids will never move from Windows, let alone Linux anyway.
Is not the "first dance" with a graphical installer. A couple a years ago there was also a tentative to implement a GUI installer (finstall).

ndisks.png


As long as we have the option to choose like in Debian GNU/Linux which installer to use, I have nothing against a graphical installer.
 
Is not the "first dance" with a graphical installer. A couple a years ago there was also a tentative to implement a GUI installer (finstall).
It doesn't mention it on the wifi but wasn't that more a project from the PC-BSD guys?

As long as we have the option to choose like in Debian GNU/Linux which installer to use, I have nothing against a graphical installer.
Other than the increased install media size, I am not against an optional GUI installer either. However Debian packages are a mess of compromises and weird scripts, so possibly isn't the best to aim for. Having a defined base in FreeBSD might make the difference.
It would be weird to install a test-based OS using a graphical installer however. That alone feels broken as a concept.

Still won't attract new users however. They are still happy with the Windows GUI installer.
 
It doesn't mention it on the wifi but wasn't that more a project from the PC-BSD guys?
No, was a FreeBSD project, not PC-BSD. They had their own installer.
Other than the increased install media size, I am not against an optional GUI installer either. However Debian packages are a mess of compromises and weird scripts, so possibly isn't the best to aim for. Having a defined base in FreeBSD might make the difference.
It would be weird to install a test-based OS using a graphical installer however. That alone feels broken as a concept.

Still won't attract new users however. They are still happy with the Windows GUI installer.
I'm not a fan of Debian either, I've just given it as an example that they offer the user the option to chose.
 
aaaand here's the first exploit:
View: https://twitter.com/hackerfantastic/status/1785495587514638559


It's incredible linux folks still blindly accept any code from that clown...
I expected nothing less than that given the fact, that it's from Poettering and relies on systemd. I mean with that stellar track record both do have in terms of security it was bound to break... and the whole thread just reads like Poettering delivered big time, again.

I ask myself the question why his software is so popular among distributions. Eg systemd & pulseaudio.
Well his software is solving real problems. Pulseaudio was released in July 2004, and alsa has been part of the kernel only for half a year or so. But merging several audio input streams from different programs into one was a real problem to solve back then, because OSS didn't do that.

The way to go for this back then was esd. It kind of did its job and was the de facto sound server standard application back then, but sucked. Most people agreed that it sucked. So when Pulseaudio came along, it sucked less (yes - unthinkable) and replaced esd.

systemd became popular because it was heavily backed by RedHat, and also because it solved real problems.

When Apple created MacOS, they've introduced launchd as its init system replacement. Solaris has SMF since ages. Ubuntu for a long time had upstart. All these systems have two things in common: 1. they are replacing INIT, 2. they try to solve the same array of problems.

Why distributions refuse to use Poettering's stuff? Well, hard to do when systemd is so overloaded with so much stuff that shouldn't be in there, and gets more bloated from release to release. Poettering is using systemd as his enabler to shittify the whole Linux world even more.

The problem with Poettering is that he thinks that he's a modern Isambard Kingdom Brunel, but fails to be as brillant as Brunel was and to deliver, too. Brunel was an engineering genius, ahead of its time, creator of many innovations like the SS Great Britain. Poettering is just selling guns which when triggered will explode.
330px-IKBrunelChains.jpg


A long time ago, I read some post or email or whatever, saying that he knew a lot about C but nothing about system adminstration. On the other hand, most distributions seem to have embraced systemd, whether because of RedHat's backing or not, I don't know.

Wrong, Poettering sucks at C. This became pretty obvious when the systemd guys tried to get dbus baked into the Linux kernel with a thing named kdbus, because they considered dbus as way too slow. So they've created a kernel component back then and handed it over for inclusion.

Linus Torvalds then first took a profiler, and looked at dbus. Then he burned kdbus, and with its his developers to ashes, because he didn't see the necessity for it.

No, I think you're right, there's the other non-potato choice: "dbus
is seriously screwed up".

That thing has almost no kernel footprint. It's spending all it's time
in user space overhead.

Quite frankly, the whole "kdbus is important for performance" seems to
be *totally* invalidated by even a minimal look at profiles for that
thing. Here's the top-15 offender list:

2.62% gdbus libc-2.20.so [.] _int_malloc
2.43% gdbus libc-2.20.so [.] free
2.31% server libc-2.20.so [.] free
2.12% gdbus libc-2.20.so [.] malloc
1.77% gdbus libglib-2.0.so.0.4200.2 [.] g_utf8_validate
1.43% gdbus libglib-2.0.so.0.4200.2 [.] g_slice_alloc
1.41% gdbus libglib-2.0.so.0.4200.2 [.] g_hash_table_lookup
1.28% server libc-2.20.so [.] _int_malloc
1.27% gdbus libglib-2.0.so.0.4200.2 [.] g_mutex_lock
1.22% gdbus libglib-2.0.so.0.4200.2 [.] g_variant_unref
1.16% server libc-2.20.so [.] malloc
1.14% gdbus libglib-2.0.so.0.4200.2 [.] g_bit_lock
1.07% gdbus libglib-2.0.so.0.4200.2 [.] g_slice_free1
1.05% gdbus libglib-2.0.so.0.4200.2 [.] g_bit_unlock
1.01% gdbus libglib-2.0.so.0.4200.2 [.] g_mutex_unlock

there's not a kernel function in sight in the top-15, and it's all
just overhead. The above is from the server side, but the client looks
similar.

If somebody wants to speed up dbus, they should likely look at the
user-space code, not the kernel side.

My guess is that pretty much the entirely of the quoted kdbus
"speedup" isn't because it speeds up any kernel side thing, it's
because it avoids the user-space crap in the dbus server.

IOW, all the people who say that it's about avoiding context switches
are probably just full of shit. It's not about context switches, it's
about bad user-level code.

Linus
 
I have found that pipewire works well for me on FreeBSD when it comes to switching between a laptop's speakers or the TV if connected with an HDMI cable. with various Linuxes (Linii?) it's been less consistent. Fedora, if I remember correctly, did it with no effort on my part. (With FreeBSD, I made a simple script to run if I'm connecting to the TV to change hw.snd.default_unit.) Void Linux worked with a little tinkering, but recently stopped being able to switch to the TV speakers, though I have sound with pipewire through the laptop speakers. LinuxMint works doing the switching but that has pulseaudio and I use pavucontrol for the switching. vermaden has an article on using it to switch around FreeBSD's sound output too, at https://vermaden.wordpress.com/2024/01/13/freebsd-desktop-part-29-configuration-audio-improvements/ but I prefer my pipewire setup.
 
Back
Top