Systemd Creator Lands At Microsoft

[poncho524]

This just strengthens my theory that Windows will eventually switch to a Linux kernel. Why develop your own when you can get one for free.

[coder]

Either way, I wish Lennart and Microsoft well. I'm personally gonna keep sticking to UNIX-like systems instead.

Huh? You know Microsoft has its own Linux distro, right?

[plan-g]

In all seriousness, I hope he can convince them to add systemd to WSL2. Having to run Windows services to manage Linux ones (as Docker Desktop does) is just so annoying.

[tildearrow]

Then hell will freeze over!

And the dreaded backslash will replace the forward slash in paths.......... -_-

[You-]

True... Pulseaudio was shit though and i'm glad to see it dying now that we have pipewire instead which is actually good. Pulseaudio was the growing pains to the pipewire we needed.

Systemd is great though, it's hated.. but only for it's monolithic structure, which is a pretty lame reason.

One thing about systemd though is that it has standardaised so many things (like puleaudio did in its field), that when a better alternative comes along, it should have an easier job in replacing systemd that systemd had in replacing all the other systems due to lack of standardisation and massive use of shell scripts.

We saw the same with pipewire replacing pulse - the standardisation, bug fixing and increase lower down the stack due to pulseaudio made the switch much easier for most use cases.

[tuxd3v]

I knew from the beginning that this guy was a trojan horse..

[coder]

This just strengthens my theory that Windows will eventually switch to a Linux kernel. Why develop your own when you can get one for free.

It's not either/or. Microsoft will probably continue using their own OS on PCs, XBox, and standalone servers, while addressing market demand for Linux in the cloud.

EEE is definitely a threat to watch out for, but probably not so easy in kernel land, where maintaining a long-lived fork is somewhat painful due to the lack of stable APIs within the kernel. However, they can start adding userspace customizations to their CBL-Mariner distro that are not GPL, and maybe pursue an EEE strategy in that way.

[coder]

In all seriousness, I hope he can convince them to add systemd to WSL2.

Perhaps that's why they wanted Lennart.

[Paradigm Shifter]

Funnily enough, when I saw the news he had "quietly left" RedHat, I was about 70/20/10 split on whether he would go to Microsoft, Facebook or Google. I weighted Microsoft so heavily because I think his "not my problem, won't fix", "not a bug" (etc) attitude would go poorly at Facebook and extremely poorly at Google.

BTW, I've long felt Lennart's design philosophy was more in line with that of Microsoft's. Seems like a good fit, in those terms.

"My way or the highway"?

The Devuanites will never let us hear the end of this.

When you've got a point, you've got a point.

Enbrace, extend, extinguish. Basically look at the .doc and .docx document standards and you will see where Microsoft employed it best.

It's not exclusively a Microsoft thing - companies have done it for years - apply pressure to a competitor, whether through pricing, advertising or other means, then offer them something, and bring the hammer down. It's a business version of survival of the fittest.

Systemd is great though, it's hated.. but only for it's monolithic structure, which is a pretty lame reason, I mean people just don't like it on a philosophical level... That said there's some truth to the argument that so many linux distributions becoming dependent on it is less than ideal, it would be nice to have some alternatives, and now that this has happened, I'm sure that we'll start seeing more distros moving away from systemd purely due to mistrust towards microsoft; and that's a good thing, cuz variety and choice is nice.

I use systemd distros (but I also use non-systemd distros, so there we are...) but I find myself ambivalent toward it because it's like a strange hybrid of parasite and cancer. Other things in the system become dependent on it, while the scope of it expands to slowly engulf the host. From a replacement init system, it now has an awful lot of functions which are not strictly init related, and only continues to grow. Some of those are convenient and useful, but it still violates the basic Unix philosophy of "write a program to do one thing and do it well".

This just strengthens my theory that Windows will eventually switch to a Linux kernel. Why develop your own when you can get one for free.

I wonder how successfully Microsoft would be able to subvert the GPL licenses? By that, I mean, argue in court that changes they've made to (originally) GPL code have altered it so fundamentally that to release it to the public would harm their IP.

And the dreaded backslash will replace the forward slash in paths.......... -_-

Please, God, no.

[coder]

"My way or the highway"?

No, not personality-related. Just in line with the overall sense I got that he's been trying to replicate the Windows userspace services on Linux.

it's like a strange hybrid of parasite and cancer. Other things in the system become dependent on it, while the scope of it expands to slowly engulf the host. From a replacement init system, it now has an awful lot of functions which are not strictly init related, and only continues to grow. Some of those are convenient and useful, but it still violates the basic Unix philosophy of "write a program to do one thing and do it well".

Exactly. Systemd does solve real problems and add real value, which makes it hard to resist. But, it slowly consumes all the Linux userspace services. One potential issue is that it lacks sufficient decoupling that you can't easily mix & match different versions of different services.

I wonder how successfully Microsoft would be able to subvert the GPL licenses? By that, I mean, argue in court that changes they've made to (originally) GPL code have altered it so fundamentally that to release it to the public would harm their IP.

I'm not worried about that. There are plenty of ways they can "embrace and extend" Linux, outside of the kernel. They don't need to fight GPL, nor would it really serve them to do so (see my earlier point about the cost of maintaining a long-lived fork).