LeFantome

I am not defending Microsoft but I have a different take.

Microsoft has already lost a the enterprise to Linux. They know it but no longer care that much. This is because the real money is in Azure (the Cloud and “the agentic web”). Microsoft makes a tonne of money off Linux and Kubernetes in the cloud. They hope to make even more money off AI. They are ok that this stuff is all Linux based. They get plenty of lock-in from volume contracts and Azure only APIs and services (especially AI sandboxes ).

However, Microsoft knows the importance of developer mindshare and influence. It is still “developers, developers, developers”. They know they cannot really stop devs from using containers and Linux but they want devs using MS software. So, they are building Linux into the Windows desktop.

They hope, I believe, that the devs will prefer the “best of both worlds” Windows experience over the “all in on Linux only” Linux one.

In some ways, they are competing more with macOS. Devs using Linux on the server had been flocking to macOS on the desktop because it is “also UNIX” but with commercial software support and a nice UX. If Linux had won on the server, Microsoft is defending the Pro desktop.

I totally agree it is wrong. It is historical.

When Windows NT was new, they had this idea that it would be compatible with many different application ecosystems via “sub-systems”. So there were going to be many different “Windows sub-systems” for various things.

There was the “Windows sub-system for OS/2” for example. And the “Windows sub-system for POSIX”. The names still sound backwards to me but I guess it makes sense if you think “This is a Windows sub-system, which one is it?”. And if you have 50 Windows sub-systems, saying “for Windows” at the end of all of them also seems a little weird.

So that naming convention was already in place when they added support for Linux. Hence the “Windows Subsystem for Linux”.

Sure. Let’s make sure that people know what this really means though.

Microsoft cannot “undo” the current license. If such a “rug pull” happens in the future, we all retain access to the code that exists at that time including all contributions from Microsoft. We can also all continue to not only use it but contribute to it under an Open Source license and keep it a vibrant, useful project if we want. Microsoft is powerless to stop us. We could fork it then or even now without the copyright assignment requirement. We have that freedom.

What the “rug pull” allows Microsoft to do is to decide, in the future, to change their policy and to make further changes themselves and not give us access to those future changes. They have that freedom.

Again, even if Microsoft did this, we could fork and carry-on. Look at Valkey and Reddis as an example.

So, the situation is that Microsoft is Open Sourcing a bunch of work that they did. The maximum possible downside is that they could stop giving us even more in the future. Our reaction is “meh”.

What concerns us is not that Microsoft can take away our freedom. They cannot. What upsets us is that they may retain or receive freedom we do not want them to have.

That is all fine. We are all allowed to think about it as we like and I guess we al value “freedom” in different ways. Sometimes though I think people misunderstand and think somehow that all the code could be “taken back”’. It cannot. Similarly, we might worry that our freedom (even the 4 freedoms) could be lost. For this code, that is not the case.

I would like to see Flatpak ported to Windows. WSL provides everything you need to support Flatpak including a Linux kernel to run on and deep GUI and network integration.

From the point of view of a user, a Flatpak app (built for Linux) could install and run natively on Windows. Flathub could be just neither App Store.

Small app developers could choose to target Flatpak instead of Win32 and have their app run on both Windows and Linux. Only one app bundle to distribute and support.

Thank of all the applications this could bring to Linux. And, once everything runs on Linux, why use Windows?

There are free RDP implentataiions

Pretty cool to see all the X Elite (Snapdragon ARM) stuff supported out-of-the-box.

Yes please. 2013 MacBook Pro? I use a 2013 MacBook Air every day. I am sure your MacBook Pro is much better.

They are selling them. Look into ESU (Extended Security Updates).

$30 a year.

I was for many of us. So, they were not totally wrong.

Even those distros are only possible if you arbitrarily decide the firmware is not software. If you want to be more honest about having free software all way down, you have to avoid AMD and Intel CPUs at the very least and most GPUs too. And, if you are not going to do that, why fuss about the BIOS?

Unless you are using a totally Free Software stack on on Open Source CPU with an Open Source ISA, it is just a question of where you draw the line between convenience and “principles”.

There are truly Open Source RISC-V CPUs. It could be done. That is not what those “libre” distros are doing.

Since none of us are using a 100% free stack, I think distros like Debian strike the right balance between “free” and “useful”.

If you do not want to use software written by Red Hat, you have to stop using Linux. Quite frankly also much of the GNU suite such as Glibc and GCC. You would absolutely have to stop using either Xorg or Wayland. Systemd is just an example of something Red Hat created but they are massive contributors to a lot of other surf too.

I you want to avoid software written by profit motivated companies, you are down to about 15% of the open source ecosystem.

Correct.

There is also iFuse and iMobileDevice