No, don't. That seems to be what they want. It's not about running linux apps on Windows. It's about blending the two so that even Linux apps are dependent on Microsft APIs and such. What is the point of bringing native DX12 to linux if not that? BTW they floated that a while back and the kernel folks said no.

It's the "extend" phase they are in now. Once they extend linux and you jump on with that gui, then the work to extinguish the other options, or just dont care because they get to take in money for every user of MS/linux hybrid.

This is painfully obvious and a bunch of people on here are saying "STFU" unless you have proof of their intentions, which is a stupid argument given the history and current efforts to "extend".

Oh, easy: it’s so games written for DX12 on Windows—that are currently run on Linux using Wine—can have a direct, low-overhead GPU driver path, rather than Wine needing to translate DX12 calls into OpenGL/Vulkan calls first.

Of course, if you’re a game-engine developer, you know you can achieve the same by just writing your engine to target Vulkan instead of DX12.

But if you’re e.g. Steam, you’d love to see more games “automatically made” multi-platform (with comparable performance), so that you can sell them to your Linux userbase.

Microsoft themselves are—maybe surprisingly—in the same boat as Steam here. The various game studios they own publish multi-platform games. Those platforms don’t necessarily include Linux, but they do often include Android. And merging code into the upstream Linux kernel is one way to get it to appear in Android.

—————

This change would also allow the graphics of Linux programs running in a VM on Windows to be accelerated through virtualization (i.e. having the X/Wayland server on Linux act as a Windows DX12 client.) It’s analogous to the reason that some VM software offers special-purpose “host-guest filesystem” drivers to the guest, that allows the guest to just pass through filesystem requests at a high level, rather than needing a virtual block device or a networked-filesystem protocol.

This is the reason Microsoft themselves offered. I feel like there’s a lot more money on being able to easily port Windows games to Android, though.

The linux community seems to believe that the world will just naturally bias toward open systems given enough time. Maybe Vulcan will own the world in a hundred years; we'll wait it out. But there's no evidence that this is the case. People just want their problems solved; they want working games, or a freakin' powerful office suite, or whatever. That's why Linux succeeded in the server in the first place; it wasn't (mostly) its libre licensing; it was just better at a lot of things, which led software to be written for it, which entrenched its position.

We can solve problems and push for openness at the same time. Linux can have DX and HyperV and whatever, while we still make open solutions better. Vulcan may actually own the gaming world one day; its very good. But I doubt LibreOffice will; Microsoft Office is just too good. Why not try to do both?

When my employer moved from Mac servers to Debian it was mostly slower performing and less obvious with the lack of a GUI.

>> Oh, easy: it’s so games written for DX12 on Windows—that are currently run on Linux using Wine

And why does Microsoft give two shits about that?

Make no mistake: any time an Xbox title isn’t explicitly chosen to be a “console exclusive” for marketing purposes, Microsoft would love nothing more than to port it to every system imaginable. Look at what they did with Minecraft.

Microsoft proprietary APIs are not and can not be part of a free software platform, that is a contradiction. They can be part of a new MS-Linux platform masquerading as the real thing until support for the real thing dries up.

(Why? Because the original studio is busy making other new games. Or because the publisher got the game produced as a one-off work-for-hire, so the original studio has no ongoing contractual relationship for the publisher to lean on. Or, sometimes, because the studio is defunct, but customers are still interested in getting new ports of the game, so "you gotta do what you gotta do.")

Publishers can't change what engine a game is written in; and there's certainly no positive ROI in a ground-up rewrite. They just have to cope with the game's codebase as-is, relying on combining small tweaks with techniques like emulation/virtualization to get the port shipped. Pushing to get the target platform to natively support the APIs the game uses, is just another such strategy.

> Microsoft proprietary APIs

What is a "proprietary API"? APIs aren't IP. Even in Oracle vs. Google, the copyright case had to focus on plagiarism of header files, not of the API.

The moment there's more than one (popular) implementation of an API, its original creator loses de-facto control over it. Third parties interested in using that API will almost always focus on the lowest-common denominator subset of the API supported by both implementations.

Which is to say that, if anything, Linux implementing DX12 would be bad for Microsoft's control over DX12, since it means that Linux could "hold DX12 features hostage" by refusing to implement them. Just like Chrome and/or Safari are currently holding a lot of HTML5 features hostage from appearing in most web-apps, by being a hold-out on implementing them.

Oh my god, the horror. Are you suggesting that I could, I dare to think it, pay Microsoft some money for a working operating system that meets all of my needs? Someone needs to call the EU and get this shut down, this is unacceptable behavior.

Look, this fear out of the linux community is why I'm hesitant to immerse myself in it. Lets say I decide to build a linux app that is dependent on Microsoft APIs, as you fear: what's stopping you from just not using it? If Microsoft allows this, and we're legal with the licensing, what gives you the right to (1) control how I build the software I want to build, and (2) control what kinds of software my users want to use?

Well, I think if we dig deep enough, the reason is some variant of "because then developers will take the easy route and use the generally pretty good and well supported Microsoft APIs, and the open linux APIs will flounder". Which is a weird combination of 'not invented here' and self-loathing from the linux community that is startlingly unhealthy. Microsoft is making an effort toward openness and interoperability here; we can argue their intentions all day, maybe they're good or bad, but its (some members of) the linux community who's saying "no, we don't want to be open, we want to live in Linux world and pretend like no one else exists." Sounds a lot like 90s-00s Microsoft.

And what's the worst case scenario? Microsoft suddenly removes their mask and admits they were the evildoer we all feared them to be, like a scooby-doo villain? Ok? They tried pulling that with the browser ecosystem, and Office, and Java, and the world is still spinning. Linux is healthier than ever.

You are free to do that now. Maybe Linux meets your needs, or maybe Windows, or something else.

>> Lets say I decide to build a linux app that is dependent on Microsoft APIs, as you fear: what's stopping you from just not using it? If Microsoft allows this, and we're legal with the licensing, what gives you the right to (1) control how I build the software I want to build, and (2) control what kinds of software my users want to use?

Nothing gives me that right. The problem isn't about you or your users. The problem is that longer term, developers like that end up using MS APIs on Linux and then all Linux users are forced to use MS APIs and we all end up paying for it. I don't pay MS anything these days, I use Apple and Linux systems and I don't want MS and other developers to ruin one of my options.

Fortunately the Linux Kernel folks are well aware of this type of thing and don't allow GPL licensed shims that serve no purpose other than supporting proprietary binary blobs.

M$ knows this.

It's time the Linux community starts to realise it too.

It's even more stupid when you consider that microsoft is going in the data collection business like facebook/google/etc. Not just their ever expanding "telemetry", taking more and more control away from the users but their acquisitions like linkedin, github, etc ( which are fundamentally data extraction companies ).

The amount of "love" microsoft/bill gates/etc gets here is rather disappointing. I'm sure it's part paid PR and partly people making a living on the windows/microsoft stack, but still sad.

Billionaires are evil, but don't you dare say anything bad about saint gates. He is here to save the world. Eerie. Facebook, google, etc are evil and people should stop using it, but I need microsoft in my life. Strange.

No, The OS is not their main money maker anymore. Even the team designing the windows OS has become much smaller ( inside sources). The focus is on the money maker - Cloud.

You can't also have Omnigraffle for Windows, yet nobody would say Windows doesn't work for that reason.

If computers no longer ran Excel, they would buy a dedicated Excel Machine to do it, and put it on their desk next to their computer.

You can make a similar argument about Premiere and After Effects, although the fact that Resolve runs on Linux is helping the situation a little.

The Autodesk products are another case.

I had a client that suffered an excel/windows ransomware attack. It started with a macro or something in an excel document that, for some reason, everybody seemed to open. My inbox was flooded with "do no open anything" emails. Then the CEO called me to check on the data I had stored on my systems. "Remember when I said I didn't use windows? Your data, at least the copy I have, is safe."

I have a working Linux desktop and I also use spreadsheets in my day to day work.

People who need Excel don't refuse to use Linux because they take issue with its stance of not-having-Excel, they don't use Linux because it doesn't have Excel.

Bad example. I've brought LibreOffice to its knees after importing a CSV file: went to adjust a column width and it became non-responsive. And this was within the last year. They have done a lot of performance work on LibreOffice and I appreciate that, but it is a long way away from Excel for serious work.

My favorite snag was on a plot with the X-axis labeled with date text at an angle. I could have horribly aliased ugly fast text, or I could have nicely anti-aliased text that slowed moving around in the sheet to a crawl. But I couldn't have nice text on a spreadsheet I could still work in.

I've crashed Excel (2016 for Mac) trying to preview CSV I was about to import.

To Microsoft's credit, they eventually fixed it.

What still doesn't work (in Mac version) is ODBC driver for Postgresql (and maybe for other databases; it is shipped with MSSQL driver only). It will also crash Excel. So to work with psql data I have to export them into file and then import into Excel. I cannot have a query saved in the spreadsheet and just refresh it.

Most people don't use Excel for "serious work". Most people use Excel to make colorful grids of text and some calculated numbers that fit in a common corporate issue 8GB computer.

Anybody who had, as their day job, working with spreadsheets would never do it with Libre Office unless they weren't really using them with a lot of data and complexity, or (B) had a bit of a masochistic streak. (Or, I guess (C), were completely dedicated to open software - which is a fair reason I guess).

The two things that will permanently hold back Linux on the desktop (And, remember, I say this as someone who has only worked on Linux for 3+ years now) - is (A) The amazing number of inconsistencies on the desktop environment (Depending on the application, I have 3 separate ways I need to initiate a "copy" - Ctrl+C, Ctrl+Shift+C, Ctrl+Right-Click+Select copy.) and (B) The lack of most decent desktop software (Chrome, Slack excepted - they are pretty much flawless on Linux).

Bringing this back to the original thread - what I would love from Microsoft, is to make Linux a first-rate citizen on Windows. Right now it's 80% of the way their with WSL. Still missing a bunch of networking stuff, and still doesn't run cron/init and friends. But It's getting closer.

Then, we have the best of both worlds - I can do all my work in Linux, and still have access to the rich desktop environment of Windows. So - count me as one person who is really happy that Microsoft is putting effort into advancing their Linux offerings by submitting these Hyper-V support patches into the Linux Kernel.

However, I feel WSL means using Linux to enhance a product (Windows) while at the same time preventing the growth of that same system in its standalone fashion. Because WSL means you won't be using a lot of the usual Linux desktop stuff, you'll be a pure Windows user in a lot of statistical data, and you won't be helping to make Linux a better OS. Not to say that it can get worse once Microsoft starts releasing Windows-only Linux components for WSL, as (I think) started happening already.

Remember - I use Linux all day for a reason - it's the better tool for the job I do.

But - Linux is incredibly lacking in highly-polished commercial software. I miss the dozens of beautifully crafted packages on OS X, and their somewhat less crafted variants on Windows constantly. That final finish, and careful craftsmanship that occurs from making every button, every font, every round-rec flawless - is something that, for the most parts - is completely absent on Linux.

I am a contrarian when it comes to Microsoft and Linux - I genuinely believe that they are going to do the right thing (And I say this as a Former Netscape Employee who knows all to well their EEE pattern) - particularly if we get to use the full desktop experience in WSL - then we'll know they are really on the right path.

But - by and large, I'm guessing we agree way more than disagree on stuff.

Vendor lock-in, OTOH, would be a valid criticism. But again, this is not that.

And the VMs are nicely contained, disk volumes can have copy-on-write snapshots, and the volumes can even be made immutable - you shut the VM down and, when you restart, it's back to its immutable state. This is becoming less than practical thanks to various self-updating things - you need to make them mutable, let the self-updates happen and then you can make them immutable again. Otherwise they'll self update a lot on every boot.

not the same but you have Office 365 that does a reasonable job for most documents out there.

On the other hand, not having Office is kind of a feature, it makes you learn better tools :)

Replace “Office” with Photoshop, Autodesk, Altium, etc. and it becomes very clear why this isn’t true. The Big Guys(tm) aren’t popular because they’re Windows and macOS programs; They’re popular because they do practically everything and no alternative comes close.

Sure, there is the bit about market dominance which leads to people learning just that tool which leads to further dominance (this is true of Photoshop especially). There’s also the bit about tools not getting better if no one uses them. But one would be hard pressed to find programs as capable as those mentioned earlier.

I think this kind of attitude is why Linux on the desktop will probably never be mainstream.

It does not need to be mainstream to be useful.

- in Excel, you can clear conditional formatting rules but you can't manage them. And text in text boxes turn into images that can't be edited, only moved.

- in Word, formulas turn to images. Complex tables break.

- Visio... well, there is no web app version of Visio.

My exotic hardware can only run on linux (wifi hacking dongles, SDR bits, very old data devices). One of the big reasons I switched to linux as a kid was so that I could play around with tools that windows hates, like enabling monitor mode on a wifi card.

Never have I seen lack of hardware support on windows. I think that’s the essence of “just works” here.

I run into some cases of software refusing to run on Linux (though curiously, I switched to Linux due to software refusing to work on Windows).

I run into wifi/printer issues (but both on Linux and Windows) and I managed Linux to Work, while Windows remained resistant to debugging and kept failing with contentless errors.

Its all kind of small things like this that make Linux grind just a little bit more for me than Windows, not saying Windows doesn't have its issues though, just a lot less.

Ever since the 2004 windows update, the dGPU in my windows laptop randomly turns on after sleep. If I try to force it off, it bluescreens. Battery life has gone from 10 hours to 3.

Tell me again about how good Windows' drivers are?

I can. I have a laptop with a high-DPI screen. It can output to a second monitor. Setting that second monitor up is relatively easy (though not as simple as it is on Windows or Mac), though getting the desktop to play nicely with a high-DPI screen and a non-highDPI screen doesn't just work. It basically becomes unusable at that point. The laptop was sold with Linux installed by default.

> but the truth is that you can run into driver problems and issues too when running operating systems you pay money for, like Windows.

Yes. Though, I haven't had that problem in years. And when I last had that problem, my solution was literally just upgrading the driver.

Listen, I've hand coded XFree86 config files back in the day to get three monitors up and running on a Slackware system. It's gotten better. But it doesn't "just work."

I think there's something to be said for the consistent experience that Apple and Microsoft achieve by being the arbiter of their entire stack.

I switched from macOS to Linux. GNOME and others are pretty good these days. I did try Windows + WSL2 out of curiosity, but my brain is just fundamentally incompatible with non-unix systems. Moreover, I was really annoyed that when Windows installed drivers automatically, it also installed a lot of Realtek and Intel crapware with it. Why?

It would be really nice though if Microsoft could make a Linux version of Office (even if it was just Wine-based), because the web version is too limited. Oh, and the Affinity suite would be nice as well ;).

A minimal version of Windows along with a good package manager and a UX closer to Windows 2000 (or at least without the whole Control Panel / Settings App confusion) would be amazing.

I have to say I find the settings / control panel dichotomy issue overblown. I rarely need to go into control panel anymore, and pretty much never without it being a clickthrough from the new settings app.

Agreed. The task bar search finds the setting I'm looking for about 99% of the time.

I could probably set up a custom image but I prefer to get the most recent build when setting up and would appreciate it being as clean and minimalistic as a fresh install of macOS (yes there are some things here too, but you have to look for it) or Linux. The plain Windows setup feels like how preinstalled systems used to be with all sorts of "helpful" apps that automatically start and make the system look "nice".

I think macOS does way better here with a minimal use of white space because I can easily have a browser open on the side to look up what I'm trying to change - though some things do appear a bit old school (e.g. networking).

You’d have to ask the driver developers. In my experience though, that crapware is the only way to configure the hardware since config files aren’t a thing in Windows.

I think there are pros and cons. PulseAudio could possibly be half the incomprehensible monstrosity that it is if it didn't have to take over the responsibilities of every audio driver out there. On the other hand, Realtek will arbitrarily disable/hide features with no way (that I've found) to do anything about as a user.

Technically, they could make these features separate and configurable from Windows sound effects options for the recording or playback device...

I've migrated from MBP to Win10 WSL three months ago and it's like a breath of fresh air for general $DayJob backend CRUD-like work. I'd still prefer a Linux laptop, but WSL is so much closer to a real Linux box.

I don't think I touched PowerShell or any .BAT stuff even once.

Virtual box is a better experience if you wanna run Linux on windows IMO.

The work in TFA describes patches to make the Linux kernel communicate with the hypervisor (sharing information like used memory pages). In my book that's pretty sophisticated.

EDIT: The work appears to be available on Github https://github.com/microsoft/WSL2-Linux-Kernel/commits/linux...

Also, Windows has decent X11 servers that work well with retina displays, which can't be said of XQuartz.

Many times in WSL2 I've had a git rebase fail because it couldn't overwrite existing files, and every time it could be fixed by just running git rebase --continue. The actual problem seems to be that WSL2 system calls return before the NTFS operation completes, so quick successions of create-after-unlink or write-after-rename can have unexpected results.

For this reason I still keep a WSL1 environment, it doesn't have the same issues (and doesn't appear to be any slower for filesystem operations).

It's definitely a great deal slower for filesystem operations inside the Linux system. Filesystem operations outside the VM's disk image go through 9p, which gives you basically the same performance for both.

Ack on XQuartz being crap (thought it does have Retina support these days, at least for non-rootless mode), but this is not really a problem of the VM.

Example: I type code ./somefolder in WSL and it opens VSCode on Windows, editing the folder I specified.

Edit: danieldk replied just before me, but I think my answer adds to that instead of just duplicating danieldks.

Also danieldks answer highlights something important: it doesn't seem to be just magic shell wrappers (or if it is they are truly magic).

VSCode is somewhat special when it comes to editing files, Notepad is not.

Whoops, typo. Windows programs from Linux, I'll fix that.

You can mount other host filesystems from inside the VM in any hypervisor I know of.

In most hypervisors you can't do a

  notepad.exe foo.txt

  TextEdit.app foo.txt

Isn't that just a binary loader, like you can do with wine/qemu on any Linux? Surely it's mostly just proxying the pipes, replicating the working directory, and maybe keeping some of the environment variables. The filesystem integration seems like the trickiest part of that altogether.

AFAIK, there's no way to achieve anything like this on macOS.

But in WSL2 the whole thing runs in a Hyper-V VM, with a customised Linux kernel (available via http://linux.microsoft.com) to help with interoperability between WSL and Windows.

It's also very paravirtualized, which is why HyperV enlightenments are required for WSL2 guest.

If you have to develop for Windows then that's of course another matter.

Of course I keep a few separate Windows PCs around to do things that Windows can only do (when I need to do them, which hasn't happened for quite some time now) like SQL Server Management Studio, Visual Studio and some games/video streaming services.

Also, I keep a Mac around to do Mac things like debugging an iOS app or helping some junior developers who only know how to use a Mac.

The only help I gave them was recommending compatible hardware. They've been using them full-time for the last year, and have been really happy with them.

Even Steam games seem to work reasonably well, which I was impressed by.

Yes, I used an anecdote, but it's an anecdote that shows a massive grain of reality. I mean, I work for a tech company full of 30-somethings and 20-somethings and only one of the engineers in a group of about 320 that I know and work with requested a Linux laptop versus Mac or Windows. If this is the case with young engineers, it's absolutely going to be even worse (for Linux) when looking at non technical older people.

No, because that's not even close to what you brought up.

> I would be surprised if 5% of computer-using 60+ year olds use Linux.

How many of those were even given the choice? I'd suspect 0%. That's the issue. The immense monoculture in multitude of places reinforces that monoculture.

Fixing that took a lot more magic than just an fsck, no matter what the incantations. Once you have trouble at that level any OS will be tricky to get going again because a lot of the underlying assumptions have failed.

Ha, yeah, that's true. I've had better luck with MacOS. I did have a total drive failure once, but the hardware had gone bad.

There is...? What sort of examples come to mind?

That consistency isn’t absolute though, rough spots like the boundary between Mach and the BSD components still exist.

However MS seems to fall back on its old bad habits and the "it-just-works" gui is less true since a while now (I would say a year): we had some major pains after every Windows update at work to the point where the admin blocked them in the firewall.

I think their release schedule (twice a year) is way too fast, and they should focus more on stability.

I still haven't managed to convert all our workstations from Win 7 to Win 10 (and yes we still have a couple XPs), there are special apps on them that needs the intervention of one of our provider, it's complicated.

Last year one of our provider sent us some machines with LTSB 2016 installed. They're already at their EOL.

Edge is not available and there's not even an image viewer on this OS, MS Photos is impossible to install (and is crap anyway, it regularly fails to show an image that any other viewer including Paint can open etc).

Color me unimpressed.

LTSB is amazing fit for things like kiosks, POS, appliance that are designed for a single purpose only. Uptime & Stability is the purpose, not enduser focus. You should never use it outside of a large enterprise where you can support it.

LTBS is meant for stuff like ATM or POS as replacement for the older Windows embedded versions.

Apple has been consistently removing native *nix tools from the system with every iteration. I was recently surprised to see it still had SFTP built-in (although I'm not sure how long that would last).

But, homebrew still does save the day. I wonder whether binary notarization requirements will affect that too.

Windows 7 did make a lot of sense, but configuring Windows 10 has become confusing. You have two versions of the configuration tools, the legacy one and the new HTML-like one, and the latter seems to be almost unusable because of how it is structured. Not to mention that the HTML-variant doesn't work when you VNC into a laptop which has its lid closed and the attached monitor is turned off. Then it's not possible to scroll, because the pages are broken, as if the HTML-renderer is unable to obtain proper display metrics.

Had Microsoft kept POSIX support on par with Win32 and Linux would never taken over.

People - users, developers, artists or anyone, should see this beyond software. They should see the philosophy that drives it, the "Free Software" movement paved way for an ecosystem where "Knowledge Freedom" was more important.