top | item 21003446

Windows 11 could run on Linux

25 points| ggm | 6 years ago |computerworld.com | reply

64 comments

order
[+] aasasd|6 years ago|reply
I'll start taking these delusional ravings remotely seriously when authors try to analyze what it would be like to implement the driver stack, the graphics stack, the security model etc. on top of a different kernel, and demonstrate that it would change anything for the best. And not just throw away compatibility with twenty years of accumulated software and hardware.

So far it's only “MS did a Linux-to-Win layer with questionable performance, it means they could totally do the opposite with negligible loss while keeping drivers for millions of device models working! And compatibility with software that hooks at every point in the Windows APIs! See, Wine does a small portion of that just fiiiine after the twenty six years of development.”

[+] thrower123|6 years ago|reply
This sort of thing reminds me of one of my favorite Microsoft UserVoices, which is asking for a Linux version of Visual Studio - hundreds and hundreds of comments from people who are presumably software developers, who seem to think that it would be trivial to port that behemoth of C/C++/.NET/whatever code, with its decades of accumulated cruft, over to run natively on another operating system.

devenv.exe isn't even 64-bit yet...

[+] nguoi|6 years ago|reply
I'm open to the concept but not where this article takes it.

Registry backups shouldn't be expected to be any easier with a different kernel. In fact, if anything it will introduce a lot more bugs, even just from programs that rely on kernel bugs.

There's no reason to believe NT is any less 'fresh' than Linux. It undoubtedly does some things better. Even as a Linux user, throwing that away would seem a massive shame.

While basing Windows on Linux may make some development easier, it also throws away the unique proposition that Windows provides.

It would take a lot of internal motivation to overcome the people who founded their careers on building an operating system for Microsoft.

[+] basch|6 years ago|reply
Less so on desktop, but Ive been arguing for a while that Microsoft could ship the Windows shell+Bing/OfficeGraph on top of Android and instantly reenter the Mobile Phone arena. It's something they could ship that installs on existing phones. People could "convert" their Android experience and "fall in love with Windows all over again" or whatever BS marketing wants to put out. The important part though would be a unique visual identity, and pushing people towards Microsoft apps and services, away from Google.

Google made this play easy by splitting off Play Services. Now all Microsoft has to do is essentially replace Play Services with Windows Services, Microsoft Services, Azure Services, or if Microsoft names it, Microsoft Windows Azure Team Site Foundations Forefront Services.

I know they have the Launcher, but to actually put a dent in Google Android, Microsoft needs to play more like Amazon and Fire os, including their own curated store.

Microsoft has embraced Android, if it Extends it, it would be logical to think that Microsoft Services would end up being a single package of cloud services that transcend Android and Windows, with the same servers, and Client implementations that best fit the OS capabilities.

[+] techntoke|6 years ago|reply
> There's no reason to believe NT is any less 'fresh' than Linux. It undoubtedly does some things better. Even as a Linux user, throwing that away would seem a massive shame.

> While basing Windows on Linux may make some development easier, it also throws away the unique proposition that Windows provides.

Like what exactly? Do you have examples or just feel this way?

[+] ecmascript|6 years ago|reply
I use Linux everyday and I don't have Windows installed anymore.

But unlike many I actually like Windows though and used to be heavily invested as a .NET developer. I switched all that away though just because I didn't like the ads, forced updates, the cloud integrations and the tracking that were included in Windows 10 and felt forced.

This is what makes Windows 10 and Windows in general bad. Not the kernel or the other things in Windows. If Microsoft would start valuing these things again Windows would be less bad in my opinion.

I just use Linux because it's better, but it's only better right now and that could change in the future. I don't believe Windows 11 would be better if it ran Linux if it still had all those privacy-issues. It would be still as shitty in my view.

I am now heavily invested in Linux and has completely dropped my investment in the Microsoft-ecosystem. I start all new projects on Gitlab because I am actually mostly tired of big american companies that track and store everything and doesn't care about you.

[+] replete|6 years ago|reply
Kinda garbage article offering nothing. Is Windows NT kernel really so bad?
[+] fuu_dev|6 years ago|reply
The NT kernel is quite good.

The reality is that there are not a lot of ppl with knowlage of the kernel and linux enthusiasts often tends to have a very baised opinion towards windows.

[+] techntoke|6 years ago|reply
Yes, the registry is terrible and Windows applications basically do whatever they want. Once you install a few applications and then try removing them you'll end up with a ton of junk left over. Dot files and standard /etc config files are a blessing. Linux package managers are orders of magnitude faster than chocolatey, which isn't part of the OS. Windows Updates are notoriously bad.
[+] AstralStorm|6 years ago|reply
It is designed for a simpler age and has not caught up. Lots of overhead for multitasking and multithreading (esp. creation), bad handling of NUMA, no actual handling of asymmetric processors (big LITTLE, per unit clock scaling). No way to enforce CPU limits per task. Expensive locking primitives with no handling of priority inversion.

And more...

[+] vallismortis|6 years ago|reply
Ok, you're crazy. The GPL licensing issues alone would be a nightmare for them, and possibly unsolvable from their business perspective.
[+] felixguendling|6 years ago|reply
Playstation, Mac OS and many others are based on BSD for that reason (and probably other reasons as well).

I cannot imagine a BSD-based MS Windows but it would be cool!

This would probably solve slow filesystem I/O (which is one of the major pain points for me when working under Windows, especially if you have a large number of small files).

However, it's important for Windows to keep backward compatibility for applications using old APIs. I think that's a major selling point vs. other platforms.

[+] dagw|6 years ago|reply
That article makes no sense. Even if we take it as given that the "Desktop Windows has had so many problems", none of those problems have anything to do with the NT kernel. Swapping one kernel for another thus have no effect on the desktop or any its "problems".
[+] crispinb|6 years ago|reply
That was my first thought too, but I do wonder about one aspect here. Doesn't Windows' relatively slow file i/o result largely from the architecture of the kernel-mode I/O subsystem? That's the impression I got from reading up on the issues Microsoft was having trying to speed up WSL (and hence, in part, the move to WSL2).

Having said that, slow file i/o probably wouldn't rank that high in the average user's list of Windows issues.

[+] kstenerud|6 years ago|reply
The layers of the os are a lot more tightly coupled than meets the eye. This would be a HUGE effort with the result of keeping the status quo. Not a smart investment. It would be far easier to port the various subsystems and grow communities around them, as they have been doing in recent years. Eventually, you'll be able to develop any Microsoft tech in Linux or windows, and then windows will slowly fade into the sunset, as the desktop isn't important anymore.
[+] simonblack|6 years ago|reply
That's not bad.

It's only taken 10 years and more since I suggested on Mini-Microsoft that Microsoft do 'an Apple' and put a Windows GUI on a Linux foundation just like Apple had put the Apple GUI on a BSD foundation. I remember that suggestion being howled down derisively by the Softies.

You can always trust Microsoft to do the smart thing .... after they've repeatedly wasted years and years and billions of dollars doing the dumb things.

[+] Spare_account|6 years ago|reply
>As it prepares Windows 11, Microsoft has been laying the groundwork for such a radical release.

Can I read about this anywhere? I didn't realise there will be a Windows 11

[+] dagw|6 years ago|reply
Microsoft has repeatedly said that there won't be.
[+] js3915|6 years ago|reply
The article it explained. And its an obvious pipedream.

Windows has all the API, wine doesn't have the API they reverse-engineered the compatibility layer.

Since Windows has the code it wouldn't take a lot. Plus a lot of drivers and what not is already implemented in linux

Biggest probably would be graphics but then again they have the API/code as they made it so they can easily connect the dots to make it work

[+] avyeed_desa|6 years ago|reply
If this really takes shape in some form I'd be interested in the status of WINE or how they would come up with any form of legacy emulation of older Windows programs. Right now it seems easier for Microsoft to emulate Linux programs with WSL or WSL2 than rewriting the Windows Kernel.
[+] samsari|6 years ago|reply
This is not a new fantasy, I've heard exactly this same theory mooted several years ago. It's obviously a very compelling one. Who wouldn't love it if Microsoft did an Apple and wrote the userspace on top of the Linux kernel?
[+] dagw|6 years ago|reply
Who wouldn't love it if Microsoft did an Apple and wrote the userspace on top of the Linux kernel?

But the whole conceit of the article is the Microsoft sucks at userspace.

[+] b34r|6 years ago|reply
Never gonna happen
[+] srbby|6 years ago|reply
Too much fluff but nothing specific. What does it mean for "Windows 11 to run on Linux"? Port Win32 and DirectX to Linux? Start supporting Wine in some way?
[+] 29athrowaway|6 years ago|reply
Major distros are shipping with a better desktop experience than Windows right now, and it has been like that for a while.

If you don't like LibreOffice, try SoftMaker FreeOffice.

[+] fiala__|6 years ago|reply
> a better desktop experience

I think that's debatable at best. Of course the experience is better because you don't get in-system ads, forceful cloud integration, and all the bad stuff that comes with Microsoft.

But my personal experience of Ubuntu is a constant struggle against randomly occurring bugs, inexplicable performance drops, crashes, and bad design. It's still worth it because it's Ubuntu, but I wouldn't celebrate the UX too much.

Out of the box, Windows is orders of magnitude easier to use for non-technical users.

[+] poooogles|6 years ago|reply
>Major distros are shipping with a better desktop experience than Windows right now, and it has been like that for a while.

Can you provide some examples here? I'm running Kubuntu and it's OK but not Windows level of slick.

Problems I'm having are display scaling and multi monitor support for different display resolutions, shitty wifi drivers and networking needing a restart whenever coming out of sleep.

The only OS that's doing what you're claiming at the moment is ChromeOS just because everything is so wonderfully integrated and _just works_. Even if that comes with handing your life over to Google...

[+] kinleyd|6 years ago|reply
Thanks. I've been trying to like LibreOffice for years and hadn't heard of FreeOffice. I have to try it out - and am keeping my fingers crossed.
[+] cm2187|6 years ago|reply
FreeOffice seems to be based on OpenOffice. There are keyboard shortcuts in PlanMaker that are very un-Excel like and that I think I recognise from OpenOffice.

I wonder what is their business model. There doesn't seem to be a paid option.