Both Windows 10 and Chrome OS are embracing the Linux kernel | Hacker News Reader

Both Windows 10 and Chrome OS are embracing the Linux kernel

pcworld.com

95 pointsindigodaddy98 comments

Maybe technologically, but not so much in spirit. The idea of "the year of the linux desktop", as I understand it, is that people en-mass get the benefits of free software and the free software desktop is a viable replacement for windows/macos.

In this case, though, people are still using their same proprietary software... they just have slightly easier access to the free software world. That's certainly a benefit for a lot of people, but it's possibly a detriment to desktop linux for this reason: Other than freedom/privacy reasons, why would I switch from windows if I can already run most of the linux stack in it and not give up Adobe programs, games, etc.? This removes most of the pragmatic reasons for switching, so I'm mostly left with idealistic ones.

I think the impractical goal of free software at any cost has hurt Linux. It's radical and it's a type of gatekeeping around here. There is always a post about Linux tech / adoption that deserves praise and some Stallman acolyte will chime in to remind us it's not a real accomplishment.

This is turning into a rant but has anyone checked out the approved distros on FSF? Many popular distros do not make the cut not because they include nonfree software but make it a little too easy to acquire nonfree software.

> I think the impractical goal of free software at any cost has hurt Linux. It's radical and it's a type of gatekeeping around here.

Completely disagree. The fact is, if you want to innact real change you need to be quite 'radical', because various pressure groups will inevitably weaken whatever gets done. So you starting position cannot be a mild/moderate one, otherwise you won't get anything done.

There's already plenty to choose from if the free software vision does not suit you. I don't need Linux to just be another Windows/macOS. I can just go straight for them if that's what I want, or go for one of the BSDs if I want a free as in cost option with no concern for software freedom.

But in the world, there's room (and need) for uniqueness. If everything would be the same, just with a different badge, you would not like to live in that world, you'd strive for diversity of opinions, foods, colors, books etc. Why should software be any different?

Proprietary software has a strong base, backed by some of the largest corporations on this planet. I think free software having a little corner for itself, one that is not under constant pressure to become just like the others, is warranted.

I don’t know about radical; if you’re talking about the same sorts of Linux environments I’m thinking of (e.g. a desktop running GNOME, using FOSS GUI apps), then, ignoring the kernel itself [which is a special case in many ways], their developers are—by volume of commits—mostly community contributors.

Would you expect that said community contributors would have any contribution/engagement at all, if the product they were contributing to wasn’t 1. sitting there with its source available on GitHub, in a state where you can compile and run your own copy of it from said source, and 2. accepting PRs from all comers, with project direction mostly up to either “the guy who wrote the code originally”, or “the whims of the people who do the most PRs”, such that 3. you gain in decision-making power in the project to the degree you contribute?

Once you have those three properties, you’ve essentially got FOSS. It doesn’t really matter if you charge for the binary distro (like e.g. RedHat does) if the developers behind the source code are essentially their own little monastery/guild, run by the will of the people who write the code, that can just up and strike if the product’s corporate owner does something stupid. If it isn’t FOSS, it’ll be forked or re-built from the ground up by said developers to be FOSS; not out of principle, but in order to put “the people writing the code” in power over the code project’s direction. (See e.g. LibreOffice.)

I wouldn’t say that making a specifically-FOSS piece of software was ever a “goal” of any of these projects, outside of Stallman’s GNU itself; rather, these projects being FOSS is just a sort of emergent property from the “anthropic principle of community-driven software engineering”: for software to exist, people—a lot of people—have to have had a hand in making it; and, unless they all worked for the same company, there has to have been some legal infrastructure in place to make many different parties able to push the project forward. That legal infrastructure is a license like the GPL or BSD. The projects that had that infrastructure thrived; the projects that didn’t have it—and there are many—stagnated or died.

(This effect only applies above a certain level of complexity, mind you; there are small software packages that are “free” but where nobody but their author has ever contributed to them. Compression libraries like zlib, and format libraries like libpng, tend to be like this; as do smaller library-ecosystem packages like your average NPM package is. In those cases, there isn’t much to grow, so stagnation is just fine. But, don’t be confused, stagnating still is what those packages are doing.)

I'm not against FOSS, in fact I'm extremely pro FOSS. I'm against the idea that simply using FOSS is the goal. My goal is productivity and most of the time that deals with leveraging free software which is great.

FOSS at any and all costs is radical. It's the very definition of radical. It's an absolute. It bugs me when the no true Scotsman argument is used to shame a well designed but partially closed or entirely closed software rather than blaming itself for not providing an equal user experience.

Most people just want to do work on their computers. Most don't give a shit about using their computer to push some everything-free ideology.

Most people are short sighted and reactionary.

For example people like me expressed massive problems with privacy @ facebook years before the recent "scandals" apparently "shocked" these normal computer users that just want to "work on their computers" with out caring about little things like privacy, ethics, or liberty.

I am sure you will consider me a "Stallman acolyte" even though i am not

Having said that, I think the compromises the "Open Source" community has given to commercial companies like Microsoft, Apple, Google, and Oracle have done much much much more harm to the community and goals than " Stallman acolytes" ever have.

Nor do i believe the goals of Free Software were impractical and as we move into a more connected world where privacy, censorship, and security are more important than ever due to automation, IoT and the normalization of technology the fact that more or less 4 corporations control the underlying fabric of that technology to the point were even the "open standards" are bought and paid for by them should be terrifying to everyone.

Of course I am just fear mongering Stallman acolyte right....

>"Open Source" community has given to commercial companies like Microsoft, Apple, Google, and Oracle

What are some examples of these 'compromises'. From what I have seen, all these companies publish source code and abide by copyright law for any open-source code they use in their products.

Or are you simply against the principle of for-profit companies making money from open-source?

Open Source vs Free Software is very simple

The Open Source community only cares about the Source Code, they are generally developer centrist, and generally work with "open sourcing" tools, libraries, languages and components used to build applications.

Looks on github's open source projects, what is the ratio of Full Applications, compared to libraries, tools, and components? Probably over 100 libraries, tools, and components for every 1 application

Open Source community has become a purely developer community that does not give 2 shits about the freedom of users or consumers. Most of the devs in open source today are building SaaS, PaaS and IaaS services that are closed source or non-source distributed "open source" services meaning that is perfectly legal to incorporate open source in to a SaaS product and not release the source code.

Free Software however believe all users of the software have a equal right to the source code as the developers. The recent Nest development of them closing their API is a clear example of the difference, if it was a Free Software based API I would have the ability to run my own Nest API server, but while Nest uses many open source libraries the service itself is Closed Source.

That is main compromise and divergence that happened between Free Software and Open Source.

As to making money off Free Software, I have no problems with that, it is not part of this issue. Open Source advocates like to bring that up as a red herring to divert the conversation from the primary debate. Often claiming that if they can not close their final product "they can not make any money" as an excuse why they do not support Free (Libre) software

>The recent Nest development of them closing their API is a clear example of the difference, if it was a Free Software based API I would have the ability to run my own Nest API server.

What changes in GPL / FSF licensing would enforce the changes that you want to achieve?

There is no change in licensing that could achieve those changes. Companies will just use other software/rewrite the libraries, and blacklist the license.

https://opensource.google.com/docs/using/agpl-policy/

From a certain perspective it's not compromise at all.

>Having said that, I think the compromises the "Open Source" community has given to commercial companies like Microsoft, Apple, Google, and Oracle have done much much much more harm to the community and goals than " Stallman acolytes" ever have.

So what? First, when you develop software you get to choose the best licence for you. Open source means that big companies can use your software. If you open the source under a permissive licence then it's because you wanted that to happen and didn't expect anything in exchange.

Second, users have the choice of what software they run, and they aren't going the way of Stallman acolytes. Why? Well, we all have our reasons.

But all in all everybody wins.

I guess that depends on your definition of "wins" which I am sure differs greatly from mine.

>Second, users have the choice of what software they run, and they aren't going the way of Stallman acolytes. Why?

yes they do have a choice, as it is given they are manipulated by marketing and ignorance of the topic. As recent as 2017 survey's showed most people (like 80% of the population) where still classified as technologically illiterate. Companies with m/billions of dollars for marketing generally win the consumer race, it has nothing to do with the technology.

>So what?

Do I really need to lay out all the issues that have already resulted from the undue influence of google has over the IETF and W3C, or the issues Orcale has caused due to their influence over java and their lawsuits against google, or Entire history of Microsoft and their anti-competitive practices?

Do I need to layout the dystopia that is on the horizon?

> Do I need to layout the dystopia that is on the horizon?

Yes. Not because I disagree, but because I’m not seeing it as potentially being a part of that 80%

For starters, consider that Stallman, and his kind set the goal posts far enough away that the field we get to play on is very large. Without them, there wouldnt be a field or anything to play with. We would be paying a subscription to watch it at home and it would probably have commercials.

This is a great point. Stallman and the FSF have an extremely positive impact merely by widening the Overton window.

Possibly. I do admire the way they put their foot down, but admittedly sometimes I just want a binary GPU driver without a lot of hassle. There's certainly an argument that getting most of the benefit of free software is better than "all-or-nothing" if that mostly results in "nothing". Especially if growing the user base of free software would give the all-or-nothing crowd more leverage with corporations in the future.

> I just want a binary GPU driver without a lot of hassle

Unfortunately, those two things tend to be incompatible. Kernel interfaces that a GPU driver might need to hook into can change from release to release. A solid GPU driver that is part of the kernel is what you _really_ want. The AMDGPU driver in the kernel is very good and gets better every release, and it is there substantially because AMD opened up their docs so such a driver could be developed. That's why I went with Team Red on my last GPU purchase and why I will on my next one. See also Intel.

I forgot where I read it, but from Microsoft's analytics it was found that majority of crashes were result of a driver bug (often) in the Graphics stack which was a decision driver for them to move the graphics stack into the userspace. The drivers still do have a kernel mode component but it is much smaller than it was in the XP days and the results show how (relatively) stable things have gotten since.

I’d question how well supported and stable the binary driver would be if the hardware company behind it isn’t willing to take basic steps to be acceptable to the wider community.

You would switch from Windows because the Linux kernel has become so ubiquitous that all the software you need is now developed for it. That's the future that this makes me hopeful for.

Aren't we almost there with software being developed for the web instead of Linux?

It's only my plain text editor that still runs on my desktop, but even that runs on electron, hehe

I wish electron would die a thousand deaths. Even excel, word, ppt and skyp are being reimplemented in it (or something similar) and the result is not an improvement (for the user)

This is the Open Source vs Free (libre) software debate

Today the trend is more open source, than free software, even with in linux.

Dev's and Companies want the source to be open but do not care if the end product is open. Free Software is losing

From a developer perspective, why would I develop software for Windows only, if I can develop it against Linux and have it work in Windows, traditional Linux desktops, and ChromeOS? (Games may still have trouble because I doubt that WSL2 will allow direct hardware access to the GPU). Once more desktop software is Linux compatible to take advantage of this cross-platform Linux API experience, it becomes easier for consumers to choose to forgo Windows all together.

The pragmatic reason is the idealistic reason: control. It manifests itself in things like forced Windows updates, even if you're actually using the computer at the time.

I wish that were enough, but even though having my computer restart when I don't want it to is annoying and microsoft's forced updates infuriate me, at the end of the day I use my computer to get work done and using linux would really make that a lot harder, unfortunately.

> even if you're actually using the computer at the time.

?? Windows 10 pops up a "schedule updates" dialog when there's an update to install. And when the time comes, I can still postpone it.

The issue with updates is that they sometimes need to be installed during restart, but many people (that I know) never power cycle their computer and then complain that it asked them to restart the computer while they were using it.

It could be like how the iPod through the last decade lured me into the Apple ecosystem. Initially it was exactly for having good music experience without having to commit to the Mac world. Then I got the iPod touch, iPad, even an iPhone. Finally it’s like what the hell, I might as well go full Mac with an MBP.

I find it sad that developers at BigCorps actually implement these kind of solutions, knowing that the resulting product is not what they themselves would want as a customer.

Apart from this particular issue, it reflects how a company with intelligent people "thinks", i.e. with only a small portion of its brain.

Replace “at BigCorps” with the more accurate “for BigCorps” and the answer is more apparent.

Morally abject behavior in the name of money exists since the oldest profession. However, most developers have a choice.

Why would giving up Adobe programs/games be a reason to switch before?

I don't think the reasons have changed much at all, just the execution. The above is an example of why one might dual boot, or run one OS in a VM, it's also a reason for chrome books/windows to ship a Linux kernel.

Maybe I lacked clarity making my point, what I meant was that a large reason for why I don't switch is that I need adobe programs and I like playing games. There are aspects of the linux desktop that would get me to switch though, hypothetically, if they really outweighed those two things, but now that I can have those things on Windows without giving up the programs I like I have less practical incentive to switch.

I feel like Linux itself (including the various FOSS Desktop Environments under that aegis) is a viable replacement for Windows/macOS, as is evidenced by ChromeOS, and by governments and enterprises deploying Linux on their workstations.

The problem is actually the opposite, at this point; Linux has seen tons of effort and been polished to hell, but all the FOSS apps that would be alternatives to proprietary ones haven’t seen nearly so much contribution.

Linux “as a product”—the kernel, the common libraries, plus any random distro’s bits chosen at random—probably has as many (if not more) cumulative man-hours put into it, per year, as the “products” of Windows or macOS have.

But can you say the same about, say, GIMP vs. Photoshop? About Audacity vs. Reason?

Even Blender, which sees tons of work put into it and is sort of a “reference platform” that graphics orgs use to demonstrate new tech (once they decide it’s not valuable enough any more to keep it proprietary), doesn’t have the same power as the proprietary equivalents.

But in many of these cases, I think the problem isn’t so much “FOSS vs. proprietary” as it is “every other solution, vs. the solution hacked on and extended by industry professionals and academics who are trying to get something novel done.” When there is such a solution, nobody else is going to be able to match its features, because its features are commercializations of R&D that may have not even been published outside of its bounds. (See e.g. Photoshop’s implementation of PatchMatch when it was first introduced; or hundreds of the technniques used in Pixar’s Renderman currently.)

In cases where the FOSS solution is the industry workhorse, the FOSS solution ends up being the thing that gets extended. LLVM is FOSS, and is also where research into optimizing compilers happens. R is FOSS, and is also the usual choice of platform to bind to for projects like WordNet, because they want to use R to do analysis against their project’s specialized data structures.

But that just means that for those industries, Linux is the locked-in platform, since it’s where these tools (or more specifically, their ecosystems of libraries and tooling, which may not always be as multiplatform as the runtime itself) run natively and/or best. But this is just as bad a situation, in many ways, as having the “industry workhorse” be a proprietary solution like Photoshop or Mathematica. Either way, you’ve got one system with all the effort concentrated on it, where sloppy algorithms get a pass because they’re the only implementation of said algorithm, and there’s no real competition to force the ecosystem forward. Sure, you could read the source and implement a competitor yourself—but given industry-developer network effects, few people start out hopeful for their prospects when they do so.

One of the things I've noticed about FOSS is that it tends to be very good at replicating proprietary solutions, but it very frequently stumbles when innovating on new ones. I think it's because there's a shortage of qualified designers and marketers that are interested in contributing. Or maybe it's just easier to get people on the same page when it's very clear what the thing is meant to be. Or maybe if you have a new idea, you want to try commercializing it before you give it away. Probably all three really. But TBH that's what makes me skeptical that userland FOSS is ever really going to be dominant.

The biggest successes of FOSS have pretty much all been by technical people for technical people (linux kernel, developer tools, servers, etc.). Those are great things, but it seems like things aimed at non-technical users almost always end up being pretty niche. (I will slightly disagree on Blender though, the latest release 2.8 release is shaping up to be really impressive -- I've been a long time Maya user and I'm actually considering switching)

I don't think this is a good thing at all.

The more Microsoft embraces Linux and makes it work seamlessly in Windows, the more developers are wont to use Windows, rather than Linux as their primary OS.

If not all, this will at-least chip away a good portion of existing developers and most importantly, upcoming developers.

This might lead to Linux being relegated to the background, permanently killing the Linux on Desktop dream.

This is indeed likely if windows GUI is significantly more stable, reliable and polished than the Linux GUI, and related GUI apps.

What I'm saying is - for Linux desktop to pose any real threat to windows, it needs to ship a GUI/Desktop comparable in quality, stability, app and driver ecosystem that rivals windows. That ship sailed 20 years ago, when philosophical infighting trumped user empathy.

Currently, I'm planning to refresh my desktop (5yo now) when the Zen 2 chips drop. I've been running hackintosh for the past 2 years, and been on a combination of Windows, Linux and Mac regularly for a long time. Most non-gui unix software works on Mac. Does that also detract from Linux?

The fact is, people will make pragmatic choices. MOST software these days is delivered online (public or private) via a web browser. People developing that software sometimes have to work on legacy applications they MUST run on Windows. Would you rather take Linux away as even an option? How do you expect people to move away from Windows if they can't support their legacy software? They'll just develop more Windows-only software in that case.

At LEAST this gives a clear path forward. And frankly, compared to Windows and Mac, all Linux desktop managers suck. Is say this as someone who pushes for Linux pretty hard in the workplace, has worked in linux environments and my next desktop probably will be Linux.

My current work environment has historically been Windows centric. The current project I'm on has been a hard push towards .Net Core (over Framework, which is Windows only) for cross-platform capability. I have made most of the CI/CD and local development workflows work with Docker (as a start). If Linux/Docker etc didn't work on Windows, guess what... That software would have been written Windows-only.

For most developers the choice is not between Linux and Windows but rather between macOS and Windows. Microsoft is trying to woo the devs who are getting disillusioned with Apple after its recent hardware woes.

Most devs just don't want to use Linux as their daily driver - it's too much maintenance, and version upgrades are consistently a nightmare.

You are predicting classic MS strategy, but I think since money is moving away from desktop computing, MS not going to actively pursuing the same. May be indirectly.

Looks like everyone dislikes the fact that Linux is only popular on an `non-free' Operating System;

But I consider that also as a success to Linux... as that means that people who feel like home at Windows can also develop Linux Software, and that's the big deal.

It's something like Emacs for Windows and Emacs for macOS, allowing gcc running on other (non-free) platforms, because they show other people how free software `tastes'.

Additionally, this makes Microsoft directly benefit from the Linux kernel and the software ecosystem; which will motivate Microsoft to invest resources on Linux.

> But I consider that also as a success to Linux... as that means that people who feel like home at Windows can also develop Linux Software, and that's the big deal.

The problem is that it does nothing to ensure the viability of Linux as a Desktop alternative. The more people will feel at home "in Windows" even to develop Linux headless applications, the less incentive there will be to run Linux Desktop. It might very well end up being a net negative in terms of alternatives for the end user.

It allows those who would otherwise continue to develop Windows-only to also develop and test for Linux. This does not remove the Linux option, it makes it an option. Try changing your perspective a little.

Where I work right now has in the past been Windows-only development. On my current project, most of the app not tied to a legacy system is cross-platform. If at least Docker didn't run via Windows, this would NOT have happened or even been an option. Another two projects are expressly supporting Linux deployments. Again, if Linux via windows didn't work, wouldn't have happened.

Not every person/company can just abandon all their in-house custom software that cost millions of dollars to create, and would be tens of millions to replace overnight.

What is exactly Linux Desktop? Isn't Linux Desktop = Linux Kernel + Mostly Free Software?

If you run a (free) Linux kernel and (free) Linux software, why can't that be considered as Linux Desktop?

What is the downside of an additionally-running Windows kernel?

Is it the usual everything-should-be-free-and-non-free-is-evil paradigm?

They're not just running side-by-side though; the Linux kernel is virtualized, with Windows being the ultimate authority. WSL gives you a little box of software freedom, but outside that box the usual rules apply: you're not allowed to look and you're certainly not allowed to change how it works. And Windows can collect whatever telemetry it wants on the running Linux instance, if MS chooses to do so.

>What is the downside of an additionally-running Windows kernel?

- You must pay hundreds of dollars to Microsoft

- You must tacitly agree with Microsoft's ToS/EULA

- You must endure whatever Microsoft decides you need to (surveillance, forced reboots, etc.)

In short, you're under corporate hegemony. Game-theoretically, it's pretty sad for you as the user[0]. You're shackled to this company not even due to the superiority of its products but the massive coordination problem that is that people collectively settled on an inertial default and the entire [consumer-facing side of the] industry sclerotically built up around that default.

>Is it the usual everything-should-be-free-and-non-free-is-evil paradigm?

TL;DR yes

>What is exactly Linux Desktop? Isn't Linux Desktop = Linux Kernel + Mostly Free Software?

I use tons of non-free software on Linux though (games, cloud software, etc.). To me, "Linux Desktop" means "desktop running a Linux distro".

I assume parent's saying, "If people who would otherwise be burdened with dual-booting into Linux to do something like use Bash can now use Bash in Windows or Chrome OS, they'll stay on Windows and Chrome OS instead of Linux." ... I guess that's a valid anxiety.

But if developers are encouraged by this to develop stuff that works on Linux because it'll now somehow work in Windows, to me that's a win -- Linux gets more dev-incentive.

0: https://youtu.be/pgY7qYaoMWQ

> What is exactly Linux Desktop? Isn't Linux Desktop = Linux Kernel + Mostly Free Software?

I think a decent working definition would be that the platform layer of the system is all free software (Kernel + Windowing System + Shell + Drivers, etc.), but user-space applications are whatever the user chooses.

> What is the downside of an additionally-running Windows kernel?

Well, arguably, you don't know what's in there, how it's using your network bandwidth, if it's sharing your personal data, etc. (I'm not a conspiracy theorist, obviously microsoft has a lot of eyes on them and it'd be detrimental for them to do something against user's trust, but that would be the argument)

> But I consider that also as a success to Linux... as that means that people who feel like home at Windows can also develop Linux Software, and that's the big deal.

The kind software development that WSL is allowing is already targeted towards the Linux servers. It does not increase the market share of Linux Servers at all, nor that of Windows Server. This is just an attempt by Microsoft to stem the flow of power users to Linux Desktop.

WSL is effectively trying to kill the Linux Desktop. All the major Linux desktops (GNOME, KDE, etc) are making great strides in recent times because a lot focus is shifting to them due to the paper cuts Windows 10 has. The Linux software that need improving are the desktop userlands, and predictably WSL does not help that.

For all the new and improved Microsoft "hearts" Linux talks, it just showed that it is still the same Halloween company.

The even worse part is seeing all those twitter thought leaders behaving as if this is the greatest thing since sliced bread. For me personally, I am not doubting Stallman ever again.

This is just an attempt by Microsoft to stem the flow of power users to Linux Desktop.

Did you see the flash video advertisement[0] Microsoft made for their new terminal? To me it feels more like Microsoft is trying to attract people fleeing from macOS because of all their recent mis-steps with the MBP.

[0]: https://www.youtube.com/watch?v=8gw0rXPMMPE

Those people would have landed at Linux because of bash, posix and all the unix underpinnings allowing them to preserve big parts of their workflow. My point still stands.

I predict that MS will be the ones to finally bring about a Linux desktop with mass market adoption. They'll do it by making it look like Windows, act like Windows, work with Windows files, and run Windows apps through emulation or virtualization.

It's pretty brilliant, really, because it solves both problems: the user doesn't have to change anything and the IT department gets the open source foundation they want.

As of now, Microsoft is only including Linux to get developers to start using windows for development of cloud apps and services. So Linux in windows, presently, is just the kernel and other essentials, that enable developers to develop and test command line applications.

P.S : By windows, I mean the Microsoft cloud ecosystem, not windows OS per se.

Agree with this sentiment.

Shipping a Linux compatibility layer in the form of an integrated Linux virtual machine does not in any way equate bringing about a Linux desktop to the masses.

It's shipping a Windows desktop that has a convenient way to do *NIX-style things, largely for developers who are leaving MS for Apple.

It also takes away from the Linux community and future support for desktop hardware on Linux. This is hurting Linux, not helping.

Kind of agree. It is further cementing the idea that Linux for the Desktop is not going to ever really be a thing-- however it does serve to reinforce the idea that Linux is and will continue to be the undisputed King of the cloud and internet. Just the fact that MS is putting so much effort, dollars and resources into the Linux ecosystem buttresses this idea.

It's the first step. No one has successfully deployed a Linux desktop to the mass market and I've outlined how MS could. And I think they will. It's actually a Very Good Thing for Linux because it cements its position as the underlying operating system -- something which is now required by enterprise IT.

As I said, this will result in more desktops running Linux, not fewer. I've never been a big MS fan but I have to say this is quite smart and will very likely be hugely successful.

This is going to kill Linux's long-term ability to run directly on consumer hardware if it becomes the primary way people access a Linux userspace.

Without users running Linux on current consumer hardware, its support to do so will atrophy.

What MS is doing is the inverse of what's best for the users. We should be running the Libre kernel+drivers directly on the metal, if you need Windows run that in a VM.

> I've outlined how MS could. And I think they will.

This is completely insane and not Microsofty at all, Satya or not.

Well, a few years ago I might have agreed with you, but even MS knows that Linux and open source have won the server. This is a company that open sourced .NET and sends people to Unity (not their product) for game development.

The issue is, "What do (desktop) users want?" And the answer is obviously Windows. If they can have their Windows cake and IT can eat Linux, too, what's not to like?

Yep, absolutely. It may not happen next year, but it will happen. The point at which this move became obvious is when they gave up on EdgeHTML and switched to Blink.

Just like web engines, operating systems are hard. They require a lot of experienced, highly paid engineers to keep working at the leading edge of tech. When you have a free, open source, collaboratively developed competitor at a similar or better level of tech as your product, why bother spending all that money on making your own? Just use Linux!

Basing Windows on Linux means MS gets the benefit of thousands of contributors from other companies and individual users. Their users don't care because Word and Photoshop still run fine on a light virtualization layer, and for less involved apps a MS-supported Wine will run them with 0 performance loss.

Hardware vendors will initially be unhappy, as they may have to rewrite their drivers for Linux, but in the longer term it is a great advantage to them - a single driver could power their hardware in literally every computer on earth except macs. That's an unparalled level of compatibility we've never seen before.

Windows would also run 100% flawlessly on other CPU archs. It would be trivial to port Windows to ARM and take advantage of the laptop-class ARM chips coming out recently. Old x86 programs run happily in the virtualized environment, and everything else is probably open source. (Admittedly a lot of companies will still ship closed source apps for Windows, so they will have to port in-house). I think UWP apps are already arch-independant and run on NET, which is already open source and I think runs on ARM.

I think the business case for WindowsLinux over plain Linux would still come from the NT kernel and userland, though. The VM delivered to every WindowsLinux user would still be a big blob of proprietary code, running on an OSS VM. Microsoft would be charging enterprise for a guaranteed-flawless and seamless way of running their legacy and big commercial apps on WindowsLinux. No need for a 'virtual desktop', all apps will run with no window decorations on the native desktop and be decorated natively. The VM would be entirely invisible. Perhaps even the GUI could be drawn outside the VM, with drawing/GPU commands going to virtual components inside the VM which redirect them outside.

"I think UWP apps are already arch-independant and run on NET, which is already open source and I think runs on ARM."

This isn't true, though it's a common misconception. UWP isn't tied to .NET or any VM. UWP apps can use a VM or use native code, and if they're native they're just as arch-dependent as any native app. The "universal" part comes from them only depending on a subset of Windows APIs which are present on all Windows editions.

The desktop edition of Windows already runs on ARM, regardless.

Regarding the "other CPU archs" point, Windows already runs on ARM: https://docs.microsoft.com/en-us/windows/arm/

I don't know that the NT kernel is going away ANY time soon... But, I am glad to see a working linux subsystem with support for SystemD (good or bad there) and Docker in the box. If volume/drive shares/mounts works well, I'll be very happy. Git bash's windows-isms mostly suck, and Docker's lack of real volume mounts is painful for anything i/o heavy.

It isn't the IT side, either... it lets developers WRITE, RUN and TEST Linux software, at least making it an option instead of more Windows-only.

It's like saying Ford will bring an electric truck to the mass market by making it sound like an ICE, act like an ICE, work with Diesel and run a 6-litre V12 to charge the batteries with a dynamo.

A complete and utter waste that does nothing more than destroy the environment it claims to benefit.

If you're a company like Ford with 30,000 Windows desktops, you might look at it differently.

That would be basically how The Mareix ended: Neo becoming one with agent Smith thus killing him.

Having a containerized Linux inside a non-free box provided by shareholder driven corporations is far away from the ideals and benefits I associate with GNU/Linux and FOSS. I know, this view may be too idealistic but I still feel it is not the right direction.

Now people who would otherwise be stuck writing windows-only systems have the option to use, develop against and deploy Linux based applications. I'd call that a win. More linux apps developed and migrated away from Windows-only, means Linux is a more viable option moving forward. Not all legacy software will be replaced/displaced or outright changed overnight.

The all or nothing approach does everyone a disservice. Also, Linux WMs are not nearly as good as Windows or Mac although some are close.

I'm a big fan of WSL and ChromeOS but how come nobody talks about the runaway success of OSS Unix on the desktop? That is by far the most successful at this point.

Is it though? I would imagine there are far more general ChromeOS users than Linux desktop users. As far as ChromeOS users that make use of the Linux features vs Linux Desktop users, well, yeah that might be debatable there...

Isn't this a good thing? Any step towards more Linux is a good thing. Now general Chrome OS users have access to the proper Linux ecosystem and also other ecosystems like Steam/GoG products.

Agree, it's a good thing. I wasn't implying otherwise. I was just replying to OP's contention that their was more "OSS Unix" desktop users than WSL/ChromeOS users. I guess OP meant OS X by "OSS Unix" though.

Did everybody misread? I said Unix, not Linux. Meaning OSX.

Calling macOS an "OSS Unix" on the desktop is pushing it quite a bit. And approximately every discussion about WSL includes how it's about making Windows an alternative to macOS for developers.

runaway success of oss Linux on the desktop? I'm a full time user but I can't tell you how much I hate he constant hardware issues. it's 2019 and I cannot hibernate or sleep my laptop reliably. the backlight doesn't work without fiddling with kernel Params.

> it's 2019 and I cannot hibernate or sleep my laptop reliably.

Oh, no worries, it doesn't even work properly on Windows on half of the hardware.

I've never really had Windows hibernation fail on a laptop or desktop for nearly a decade now. Inversely, Linux doesn't fail on hibernation, it's usually rock solid there, but has other problems, like battery not being picked up, or nonexistent power saving (probably because it thinks it's a desktop).

Source: I like installing Linux on obscure things like Chromebooks and Windows 2-in-1s.

I manage a crap ton of windows systems... First thing i do is disable Hibernate on all of them. It does not work very well on windows either. So that is not a linux thing. Hibernate should simply not be a power state.

Nor should sleep in IMO but I do keep that on in windows.. On and Off should be the 2 power states. People just putting their windows computers to sleep for months on end causes all kinds of weird shit to happen in windows

>but I can't tell you how much I hate he constant hardware issues

And you believe windows works perfectly on all hardware? I select my personal computers based on Linux hardware support and I have zero issues. Windows has all kind of wierd driver issues, random shit that happens with some hardware.

Every issue you can come up with that linux has a problem around, i can find the same problem in windows. Welcome to General Computing

I will suggest that Windows has far fewer problems... and while my next desktop (months away) will likely be Linux, it is definitely not nearly as friendly as either Windows or Mac for most users.

Of course, if you don't manage your own computer and mostly run out of a browser, it doesn't matter nearly as much. It's significantly better today than a decade, or even 2 years ago. About every 2-3 years, I've run Linux as my main OS for anywhere from a month to I think 4 months being the longest. It's not stuck because invariably I hit a sticking point that just gets me to say "fuck it." Running a hackintosh for all its' faults is generally easier than running Linux as a desktop OS, which is a really sad statement.

It's not that Windows doesn't have problems, it's that it has them a lot less frequently. I say this as someone who prefers to develop against Linux (mostly web ui and server backend, etc). I absolutely hate the frustrating windows-isms in Git bash. I find the WSL 1.x isn't sufficient, and the lack of properly performing volume mounts in Docker for Windows very frustrating to say the least.

I, frankly, welcome a more transparent/fast/effective/complete option for linux software development in windows. What has come so far has allowed me to push it as an option moving forward that wouldn't otherwise be there.

> People just putting their windows computers to sleep for months on end causes all kinds of weird shit to happen in windows

Not just Windows, Linux/BSDs have problems, too, because just like Windows they tend to clean up temporary files only on reboot / relogon. Though Windows in particular had a number of issues where it would fill up the entire system drive, no matter how big, if you let it run long enough (related to not cleaning up extremely verbose log files and a couple problems with Windows Update).

> Windows has all kind of wierd driver issues, random shit that happens with some hardware.

Recent 1-2 years of nVidia drivers seem to generally lack in stability, with many people reporting frequent driver resets and outright game crashes esp. on Windows 7 and 8. Linux drivers aren't in a particularly convincing state, either. Issues with dis/reconnect displays, suspend/resume issues, 3D application hangs, hardware video crashing drivers (an issue on both OSes) etc. etc.

I clearly didn't say Linux. I said Unix. OSX.

A lot of OSX is not OSS, so I wouldn’t be surprised by the confusion.

> Get this: Chromebooks also support Android apps, as Google’s mobile operating system is also built on Linux. Which means that developers could run software from three different operating systems at the same time on a Chromebook.

No, two only. ChromeOS runs on the Linux kernel and the rest was merely a thin web client. ChromeOS apps are simply offline web applications.

Crostini is The Linux subsystem/container for chrome OS. They have put a very large amount of effort to making this work well (mainly around security so Linux apps can't break the security guarantees that chrome OS provides). They are working exposing USB to it, could aid in Android development.

https://chromium.googlesource.com/chromiumos/docs/+/master/c...

https://betanews.com/2019/04/13/usb-support-linux-apps-chrom...

I'd love to see ChromiumOS expanded into a much larger general Linux platform... it has the best chance of a real Linux Desktop environment for the masses. I don't know that it will ever really happen though.

Canonical could certainly move in that direction for their chosen window manager and ui shell.

It might unify development more. Now all developers (Linux, Windows, Chrome, Mac) have access to the same command line tools.

The remaining things I wait for on windows: built in gpu pass through from windows to linux and enabling wsl on enterprise licenses by default.

I'm curious what the timeframe is for Fuschia to displace Linux in Android, and then maybe ChromeOS.

Lel zircon will never replace Linux, Google is acting retarded here, they have not the ressources to make zircon competitive with Linux.

Will Microsoft "Linux kernel" devs influence Linux roadmap?

Linux never really had a traditional roadmap, it just is the sum of all contributions, mediated by what subsystem maintainers and Linus will accept for merging. So sure, Microsoft devs will influence Linux development just like the other big contributors. Google, Oracle, RedHat, Intel, et.c.

Just tried getting docker with wsl + Ubuntu to work on windows 10 yesterday, and it was a horrible, horrible experience.

Couldn't get it to work in the end.

Made me realise how peaceful macOS with brew really is.

Even getting zsh to work with hyper properly was super painful. If you ask me Linux on windows is not usable yet.

Checkback in an year?

WSL2 should be available soon and will be a better experience... I've just been running Docker for Windows and using the bash that comes with Git for Windows. It is not great, but the best option(s) I've used so far.

edit: current Docker for Windows does not use REAL volume mounts, don't use it for any live databases in containers, it's good enough for a backup target volume, but live data is too fidgety.

Have you tried using pengwin[[1]](https://github.com/WhitewaterFoundry/Pengwin)? I got the docker client working for windows in just a few keystrokes last night.

[1] [https://github.com/WhitewaterFoundry/Pengwinhttps://](github...

It's not as seamless as docker on Mac or Docker CE on Linux, but it doesn't take much to get docker working in WSL:

https://nickjanetakis.com/blog/setting-up-docker-for-windows...

Reply on news.ycombinator.com