I’d just like to interject for a moment. What you’re refering to as Linux, is in fact, Win32/Linux, or as I’ve recently taken to calling it,
↫ The loss32 homepageloss32Win32 plus Linux. Linux is not an operating system unto itself, but rather another free component of a fully functioning system made useful by WINE, the ReactOS userland, and other vital system components comprising a full OS as defined by Microsoft.
Joking introduction aside, this is exactly what you think it is: a Linux kernel with the Windows user interface running on top through Wine. I’m sure quite a few of use mused about this very concept at some point in time, but hikari_no_yume went a step further and created this working concept. It’s rough around the edges and needs a ton of work, but I do think the idea is sound and could offer real benefits for certain types of users.
It’s definitely a more realistic idea than ReactOS, a project that’s perpetually chasing the dragon but never coming even close to catching it. Not having to recreate the entire Windows NT kernel, drivers, and subsystems, and using Linux instead, is simply a more realistic approach that could bring results within our lifetimes. The added benefit here is that this could still run Linux applications, too, of course.
hikari_no_yume is looking for help with the project, and I hope they find it. This is a great idea, with an absolutely amazing name, too.
It’s more than than, it makes use of the ReactOS userland to replace most of the GNU userland. It essentially tries to achieve the “OS personalities” vision of Workplace OS:
https://en.wikipedia.org/wiki/Workplace_OS
Amazing.
While neat, it doesn’t really change the fact ReactOS userland is still a huge project by itself, even without the kernel. We’ve already seen a similar effort by ReactOS developers with Arwinss, but that’s dead because the main project advanced enough that Arwinss didn’t provide much advantages. Also under “How do I help?” -“Making a Linux distro generally,” yeah, this guy is in over his head.
The author is a woman, not a guy btw.
dark2,
And this comes at the time Windows itself is moving towards having more Linux inside, and newer desktop applications, including the Start Menu itself is no longer a native Win32 application.
(Yes, I know it is a stupid design. But one of the reasons Windows 11’s regressions is their move to Electron, XAML, “React Native” and similar runtimes for essential parts of the desktop).
So, by the time “Loss” 32 (really?) reaches any usable status, Windows itself has already moved on.
Part of the reason for trying to recreate the kernel was to be able to use existing windows drivers – which would provide support for some hardware that linux does not work with.
And for running linux programs, that’s what WSL is for although reactos has obviously not got that far yet.
Sure, but replicating the functionality of the Windows NT kernel proved a much bigger task than anticipated, the end result being that ReactOS has a bigger problem running on modern hardware than the Linux kernel.
Ironically, when ReactOS started this was really useful because hardware manufacturers did not want to support Linux, but these days Linux kernel more and more has better driver support than the original manufacturer’s driver on Windows.
Lennie,
I still lament the existence of hardware that never gets a linux driver. But in my experience we’ve ended up having to throw away more EOL hardware because it stops working after windows upgrades. I hate having to throw away working hardware on account of windows drivers. Linux, while suffering from less official manufacturer support, typically supports hardware for the long haul.
My last encounter with this was 2025 with specialized industrial usb equipment containing prolific chips. Prolific pushes out updated drivers through windows update that officially brick the hardware on windows 11. The original windows 7 drivers are compatible with windows 11, however between new broken drivers and microsoft’s new forced update policies, prolific are able to weaponize windows updates into a fairly robust planned obsolescence mechanism for older hardware so their customers are forced to buy new hardware. I’m not naive enough to believe the government would intervene to protect consumers, but this really ought to be illegal. It’s not just about e-waste, but corporations taking specs to actively break working hardware shouldn’t be something that consumers have to tolerate.
That logo is just… *chef’s kiss*
The biggest issue with this is it needs a control panel for controlling display/sound/network settings. Ideally it would look like Windows 2000 by default and support skinning to look like XP/7/10/11. Obviously drivers would be Linux based and a Bash/ZSH Shell would still be required for some tasks.
While I don’t care about running an explorer.exe clone as my desktop, I really hope that this project can improve the non-gaming parts of WINE and make the latest versions of advanced desktop applications such as DxO PhotoLab and the Adobe suite run reliably on Linux.
Just based on the screenshot, this is absolutely what I would want for a retro-Windows inspired *nix OS, something we could have had with SerenityOS if its lead developer hadn’t abandoned the project to work on his browser and then outed himself as a transphobe and misogynist.
I’ll be following this project very closely!
Forgive my ignorance, but does this acyually offer any significant improvement over a Linux disro running WINE? I get ReactOS, but i really fail to see what this idea brings to the table.
Wouldn’t a free and open source Windows 3.1/Win95 with win16/win32 compatability be a better use of time and resources?
Linspire, Lindows, etc. History has proven making Linux look like Windows is putting lipstick on a pig. The problem is the GNU part of GNU/Linux. It’s overdue for a replacement with a real desktop instead of a piecemeal thing that has such an incredible tendency to break something that “immutable” distros to stop that have become a thing.
I don’t think that’s a particularly good argument. Surely this project will still use x11/Wayland and the Linux multimedia stack? Then you’ll need the supporting software, like GNU, which essentially makes this project a Linux distro with a WINE focus.
If the project is truly to graft the Win32 userland onto the Linux kernel, it’s going to be like trying to graft a car body to a steam locomotive chassis. All the bits are going to be in the wrong place and the plumbing won’t match at all, and you’ll spend years/decades to get something that is inferior to both.
As long as Microsoft keeps forcing end users and installers to Microsoft accounts, Ai, and Microsoft store for software, I see this and possibly ReactOS taking off.
As a software engineer and Linux user since year 2000’s, for me, the Achilles’ heel of Linux is ABI compatibility.
While for open-source software, it is OK, as distros can just rebuild everything then the binary compatibility changes due to a new lib6 or stdlibc++, for commercial apps, that is a nightmare.
Maybe just implementing Win32 API as a compatibility layer, Linux can fix that issue for good.