Hacker News Reader

Comment by elif | Hacker News Reader

Every time I praise WSL on hn I pay the karma tax but I will die on this hill. WSL is more powerful than Linux because of how easy it is to run multiple OS on the same computer simultaneously. It's as powerful as Linux with some janky custom local docker wrappers for device support, local storage mapping, and network mapping. Except it's not janky at all. It's an absolute delight to use, out of the box, on a desktop or laptop, with no configuration required.

Edit: for clarity, by "multiple OS" I mean multiple Linux versions. Like if one project has a dependency on Ubuntu22 and another is easier with Ubuntu24. You don't have to stress "do I update my OS?"

View on HN

You can accomplish the same with Distrobox on Linux, but there's definitely something to be said about having the best of both worlds by running Windows + WSL.

I honestly think Microsoft could win back some mind share from Apple if they:

* Put out a version of windows without all the crap. Call it Dev edition or something and turn off or down the telemetry, preinstalled stuff, ads, and Copilot. * Put some effort into silicon to get us hardware with no compromises like the Macbooks

I'm on Mac now, and I jump back and forth between Mac laptop and a Linux desktop. I actually prefer Windows + WSL, but ideologically I can't use it. It has potential - PowerToys is fantastic, WSL is great, I actually like PowerShell as a scripting language and the entire new PC set up can now be done with PowerShell + Winget DSC. But, I just can't tolerate the user hostile behavior from Microsoft, nor the stop the world updates that take entirely too long. They should probably do what macOS and Silverblue, etc. do and move to an immutable/read-only base and deploy image based updates instead of whatever janky patching they do now.

Plus, I can't get a laptop that's on par with my M4 Pro. The Surface Laptop 7 (the arm one) comes close, but still not good enough.

I'm not saying it's a perfect solution, but with Windows 11 Pro and group policy I was able to disable all of the annoying stuff, and because it is group policy it has persisted through several years of updates. It is annoying you have to do this, and it does take some time to get set up right. But it's a solution.

That said I'd pay for a dev edition as you described it, that would be fantastic.

You can make your own clean version, legally, with this file. https://schneegans.de/windows/unattend-generator.

I get customers and most people don't know about it but it's kind of ridiculous that techy people in a tech forum don't know how to do it.

There is no flavor of Windows 11 that is acceptable. Even the UI itself is a disaster. A cornucopia of libraries and paradigms from React Native to legacy APIs as if an interdimensional wave function of bad ideas had collapsed into an OS, but with ads.

Windows LTSC already exists, but Microsoft, in all their wisdom, restricts it to enterprise licensees only, and seems to actively discourage using it as a desktop OS. The first problem is of course fixable with some KMS server shenanigans, but the second can be kinda painful when it comes to keeping drivers up-to-date, installing apps that rely on features LTSC excludes (and doesn't provide an easy way to install manually), etc.

I've often said that if Microsoft had just iterated on Windows 2000 forever I'd probably still be a full-time Windows user. If Microsoft had maintained an LTSC-like Windows variant that was installable from the normal retail installation media and with a normal retail product key (at the very least Pro, but ideally Home), that also likely would have kept me on Windows full-time instead of switching to Linux as my daily driver.

I use Windows 11 IoT Enterprise LTSC, which as far as I'm aware has all the features that Pro has (plus the IoT Enterprise stuff) and zero bloat. I switched to it from my already de-bloated 11 Pro installation (because it removes some telemetry you're normally unable to disable) and have had 0 issues with it. I can't say I activated it using a normal retail product key, however, there are easy solutions to that.

Ya I totally get that. The way I view it is that windows is a glorified driver support layer and any actual work i do is almost exclusively in the Linux container.

When I used to have free time it was great for games too

> I can't get a laptop that's on par with my M4 Pro.

This is the only reason I have not requested a windows laptop from my company. WSL is better for docker development in basically every way than a mac can be (disclaimer: haven't tried orbstack yet, heard good things, but my base assumption is it can't be better than WSL2) except it is literally impossible to get hardware as good as the M3 or M4 for any other OS than macOS.

I replaced my m1 with a snapdragon laptop running Win11 and upgraded that to pro. For what I do with it, it runs great with very long battery times, for less than Apple quoted to repair the m1. I don't use the copilot features and haven't seen any ads so far, except maybe for office during setup.

(Used 15ys OSX, now Win11)

The biggest difference between OSX and Windows is, Apple adds (some say steal) functionality from competition, and open source. They make it neat. On windows to have something working, you need a WezTerm, Everything for search, Windhawk for a vertical taskbar on the right, Powertoys for an app starter, Folder Size for disc management etc. If you spend a lot of time, Win11 can be ok to work with.

If Powerpoint and Affinity would work on Linux, I'd use Linux though.

Maybe just for your specific preferences. Terminal is plenty fine. Vertical taskbar on the right is straight up user preference. PowerToys for an app starter? Like Alfred? The start search does a decent enough job of that. Folder Size is nice, but enumerating all files is very taxing.

Oh running Ice to wrangle the menu bar app icons or Rectangle to properly manage windows ('cause Apple screwed that one up) must be unnecessary.

Each OS is going to have extension applications to improve on the OOTB experience. This is an invalid argument to choosing one over the other.

>Windhawk for a vertical taskbar on the right

Huh? Windows supports vertical taskbar.

Outside US and countries of similar income level, Windows is doing quite alright in mindshare, and will keep doing that unless Apples stops pretending being the computer version of audiophile.

I on the other hand cannot get an affordable Mac that has the same GPU, disk space and memory size as my workstation class laptop.

To the tech savvy, there is essentially only one advantage to running Windows, and that is the ability to run Windows-only software. In all technical respects - control, performance, flexibility - it is inferior to the alternatives. Don't confuse vendor lockin with technology.

I find it dismaying that people on Hacker News willingly submit to incredibly user-hostile behavior from Microsoft and call it "the best of both worlds". Presumably a nontrivial proportion here are building the next generation of software products - and if we don't even respect ourselves, how likely is it that we will respect our users?

"I find it dismaying that people on Hacker News willingly submit to incredibly user-hostile behavior from Microsoft"

And I find it funny that the crowd that spends whole days implementing user-hostile features in yet another SaaS crapware has so much to say about Microsoft's bad behavior.

There is an additional reason: Some (many?) people simply prefer the Windows UI conventions (once you remove all the enshittifications post Windows 7).

I don't think Microsoft losing the mind share has anything to do with software. Macbooks are winning the laptop war because of superior hardware.

Only on countries where people earn salaries big enough to pay for the Apple hardware tax.

Superior hardware with terrible software. Also they straight up artificially limit their hardware so they don't cannibalize their sales, which is slightly understandable, but they do it in the dumbest ways. My SOs MacBook Air can only do one external monitor, even though it has the same specs as her work Pro. Oh and good luck actually getting that external display to work, I swear only like 50% of USB-C docks work on the platform.

>Macbooks are winning the laptop war because of superior hardware.

No. This is just you repeating marketing.

No Nvidia chip = B tier at best.

I have a $700 Asus with a 3060 that is better. Go ahead and scale up to a $2000 computer with an Nvidia chip and its so obviously better, there is nothing to debate.

No one cares about performance per watt, its like someone ran a 5k race, came in 3rd and said "Well at least I burned fewer calories than the winner!"

This would be fantastic. But Microsoft doesn't have to do this. Their users are captives.

Some of them are.

But the increasing market share of Macs and even Linux these days plus the ever increasing of OSS initiatives from Microsoft points out that Microsoft knows a lot fewer of their users are as captive as they were in the 90's, for example.

There's a dedicated settings page for quickly setting popular dev settings such as showing extensions and full paths. Getting rid of the rest just involves tweaking a few other settings like don't show tips or welcome screen. I also hide the weather and news widget because it's tabloid rubbish but many people seem to love it.

> nor the stop the world updates that take entirely too long

Interesting enough, that beyond release upgrades, happening may be once a year, all or may be 99% of updates took ~5 minutes of interruption of me, including needed reboot. I really wonder how others manage to have "entirely too long" updates.

5 minutes is too long. My Debian systems never demand that I update them. When I update them, it never even takes two minutes.

> a version of windows without all the crap

LTSC is a version like that

> "Microsoft doesn't make any release from the Long-Term Servicing Channel available for regular consumers. The company only makes it available to volume licensing customers, typically large organizations and enterprises. This means that individual users cannot purchase or download Windows 11 LTSC from Microsoft's website."

https://www.windowscentral.com/software-apps/windows-11/what...

> without all the crap

as far as MS are concerned, that crap is their business.

Or, possibly, that crap is the multitude of little software empires build by the management layer now in control..

"More powerful than Linux" is silly. It's a VM. The most useful thing is that it does a bunch of convenience features for you. I am not suggesting that it is not extremely convenient, but it's not somehow more powerful than just using Linux.

You know what's even more convenient than a VM? Not needing a VM and still having the exact same functionality. And you don't need a bunch of janky wrapper scripts, there's more than one tool that gives you essentially the same thing; I have used both Distrobox and toolbx to quickly drop into a Ubuntu or Fedora shell. It's pretty handy on NixOS if I want to test building some software in a more typical Linux environment. As a bonus, you get working hardware acceleration, graphical applications work out of the box, there is no I/O tax for going over a 9p bridge because there is no 9p bridge, and there is no weird memory balloon issues to deal with because there is no VM and there is no guest kernel.

I get that WSL is revolutionary for Windows users, but I'm sorry, the reason why there's no WSL is because on Linux we don't need to use VMs to use Linux. It's that simple...

Yeah if you are working with Linux only, its better to go full linux.

WSL2 is really handy when you want to run other software though. For example, I use Solidworks, so I need to run windows. Forscan for Ford vehicles also has to run under Windows. Having WSL2 means that I can just have one laptop and run any software that I want.

My development is mainly Windows and I prefer Linux host with Windows VM guests. The experience is more stable and I can revert to a snapshot when Windows or Microsoft product update brakes something or new test configuration does. It also allows to backup and retain multiple QA environments that are rarely used, like a client's Oracle DB. It is nice being able to save the VM state at the end of the week and shut it all down so you can start the next right where you left off. Cannot do that when your development environment is the bare metal OS. Windows has known issues of waking a sleeping laptop.

In the same spirit if "it depends", there are other options that may work for people with different Linux/Windows balance points:

* Wine is surprisingly good these days for a lot of software. If you only have an app or two that need Windows it is probably worth trying Wine to see if it meets your needs.

* Similarly, if gaming is your thing Valve has made enormous strides in getting the majority of games to work flawlessly on Linux.

* If neither of the above are good enough, dual booting is nearly painless these days, with easy setup and fast boot times across both OSes. I have grub set to boot Linux by default but give me a few seconds to pick Windows instead if I need to do one of the few things that I actually use Windows for.

Which you go for really depends on your ratio of Linux to Windows usage and whether you regularly need to mix the two.

> Forscan for Ford vehicles also has to run under Windows.

I've successfully run it with WINE. Thought, my Forscan executable was 3 years old or so and that may have changed, but I doubt it.

> For example, I use Solidworks, so I need to run windows.

Right. One of the things a lot of people don't get is the extent to which multidisciplinary workflows require Windows. This is particularly true of web-centric software engineers who simply do not have any exposure to the rest of the engineering universe.

Years ago this was the reason we had to drop using Raspberry Pi's little embedded microcontroller. The company is Linux-centric to such an extent that they simply could not comprehend how telling someone "Just switch to Linux" is in a range between impossible and nonsensical. They were, effectively, asking people to upend their PLM process just for the sake of using a little $0.50 part. You would have to do things like store entire OS images and configurations just to be able to reconstruct and maintain a design iteration from a few years ago.

WSL2 is pretty good. We still haven't fully integrated this into PLM workflows though. That said, what we've done on our machines was to install a separate SSD for WSL2. With that in place, backing-up and maintaining Linux distributions or distributions created in support of a project is much, much easier. This, effectively, in some ways, isolates WSL2 distributions from Windows. I can clone that drive and move it from a Windows 10 machine to a Windows 11 machine and life is good.

For AI workflows with NVIDIA GPU's WSL2 is less than ideal. I don't know if things have changed in this domain since I last looked. Our conclusion from a while back was that, if you have to do AI with the usual toolchains, you need to be on a machine running Linux natively rather than a VM running under Windows. It would be fantastic if this changed and one could run AI workflows on WSL2 without CUDA and other issues. Like I said, I have not checked in probably a year, maybe things are better now?

EDIT: The other reality is that one can have a nice powerful Linux machine next to the Windows box and simply SSH into it to work. Most good IDE's these days support remote development as well. If you are doing something serious, this is probably the best setup. This is what we do.

Did you know that Forscan works flawlessly under Wine if you're not using Bluetooth?

I really want to like Windows 11, and I enjoy using WSL, but Microsoft treats me too much like an adversary for me to tolerate it as a daily driver. Only a complete scumbag of a product manager would think pushing Candy Crush ads is a good idea.

I’ve got an airgapped Toughbook that I use for the few Windows apps I really need to talk to strange hardware.

I run Windows in a VM where I need windows. It’s so much easier to fix a broken Linux installation than a broken Windows installation.

My coworkers stubbornly try to use WSL instead of Linux directly. They constantly run into corner cases and waste time working around them compared to just using Linux. Some tooling detects that it is running on Windows, and some detects that it is running on Linux. In practice, it's the worst of both worlds.

Saying running full Linux avoids wasting time on fiddly workarounds kinda blows my mind.

Full hardware support is still not a given, and Windows emulation is still need for so many cases (e.g. games, specialized software etc).

Until I can choose any machine based on form factor and specs alone and just run Linux on it, WSL will the best version of Linux it can run.

What detects it is running on windows out of interest?

I use WSL extensively, with lots of languages, and I’ve never had anything do that.

It’s running in a VM, so that would be some kind of weird VM escape?

I would love to hear about these edge cases and which tooling fails to detect they they were launched form linux.

Sounds a lot like a picnic problem but you didn’t give nearly enough details.

If on WSL2, they need

[interop]

appendWindowsPath=false

section in /etc/wsl.conf.

Then everything will go flawlessly.

> "More powerful than Linux" is silly. It's a VM.

I don't think it's silly. Sure, it's a VM, but it's so nice that I barely reboot into Linux. You get the best of both worlds with WSL.

For me, the best part of running Linux as the base OS is not having to deal with Windows.

No ridiculous start menu spam; a sane, non-bloated operating system (imagine being able to update user space libraries without a reboot, due to being able to delete files that other processes still have opened!); being able to back up my data at the file level without relying on weird block-level imaging shenanigans and so much more.

How is inverting the host/guest relationship an improvement on that?

But you still get the worst of the Windows world, which is more than many are willing to deal with. I was using windows for years as my main gaming OS, but after they announced W11 being the only way forward. Switching to Linux on the desktop was like a breath of fresh air. I'll leave it at that.

If I were to run an OS on a VM it's gonna be windows, not Linux

> You get the best of both worlds with WSL.

You obviously don't. Maybe WSL is the best compromise for people who need both Windows and Linux.

But it's ridiculous to think that WSL is better than just Linux for people who don't need Windows at all. And that's kind of what the author of this thread seems to imply.

Similarly powerful would be totally fine. More powerful really is silly. Personally I couldn't make a lot of my workflows work very well with WSL2. Some of the stuff I run is very memory intensive and the behavior is pretty bad for this in WSL2. Their Wayland compositor is also pretty buggy and unpolished last I used it, and I was never able to get hardware acceleration working right even with the special drivers installed, but hopefully they've made some progress on that front.

Having Windows and Linux in the same desktop the way that WSL2 does obviously means that it does add a lot of value, but what you get in the box isn't exactly the same as the thing running natively. Rather than a strict superset or strict subset, it's a bit more like a Venn diagram of strengths.

Technically it's not a VM, it's a subsystem, the same way Win32, Win64, Posix, OS/2, etc. are.

It's a feature of the NT-family of kernels where you can create many environments sharing the same underlying executive and HAL.

It's a quite interesting way to build an OS: https://en.wikipedia.org/wiki/Architecture_of_Windows_NT

As everyone said, WSL2 is actually virtual machines and it is what most people are actually using now. That said, I feel the need to chime in and say I actually love WSL1 and I love Windows NT the kernel. It bums me out all the time that we probably won't get major portions of the NT kernel, even an out-of-date version, in some open source form.

I like Linux, and I use Linux as my daily desktop, but it's not because I think Linux or even UNIX is really that elegant. If I had to pick a favorite design it would be Windows NT for sure, even with all its warts. That said, the company behind Windows NT really likes to pile a lot of shit I hate on top of that pretty neat OS design, and now it's full of dubious practices. Automatic "malware submission" on by default, sending apps you download and compile yourself to Microsoft and even executing them in a VM. Forced updates with versions that expire. Unbelievable volumes of network traffic, exfiltrating untold amounts of data from your local machine to Microsoft. Ads and unwanted news all over the UI. Increasing insistence in using a Microsoft account. I could go on and on.

From a technical standpoint I do not think the Linux OS design is superior. I think Linux has some amazing tools and APIs. dmabufs are sweet. Namespaces and cgroups are cool. BPF and it's various integrations are borderline insane. But at its core, ... It's kinda ugly. These things don't all compose nicely and the kernel is an enormous hard-to-tame beast. Windows NT has its design warts too, all over, like the amount of involvement the kernel has in the GUI for historical reasons, and the enormous syscall surface area, and untold amounts of legacy cruft. But all in all, I think the core of what they made is really cool, the subsystems concept is super cool, and it is an OS design that has stood up well to time. I also think the PE format is better than ELF and that it is literally better for the capabilities it doesn't have w.r.t. symbols. Sure it's ugly, in part due to the COFF lineage, but it's functionally very well done IMO.

I feel the need to say this because I think I probably came off as a hater, and tbh I'm not even a hater of WSL2. It's not as cool as WSL1 and subsystems and pico processes, but it's very practical and the 9p bridge works way better than it has any right to.

Thanks for pointing this out.

It used to be. They moved to a VM.

Turns out that it's easier to emulate a CPU than syscalls. The CPU churns a lot less, too, which means that once things start working things tend to keep working.

You're thinking of the POSIX personality of Windows NT of old. This was based on Interix and has been deprecated about two decades ago and is now buried so deep that it couldn't be revived.

The new WSL1 uses kernel call translation, like Wine in reverse and WSL2 runs a full blown Linux kernel in a Hyper-V VM. To my knowledge neither of these share anything with the aforementioned POSIX subsystem.

WSL 2 is actually virtualized despite the name

WSL1 was a subsystem. WSL2 is mostly a VM.

They had to give that up because it was too slow, I think for IO. Unfortunate.

I mean... WINE does the same on windows, but microsoft refuses to release their API docs for all internal APIs. They release WSL by relying on Linux's open-ness, while refusing the same for themselves.

The integration between Windows and the WSL VM is far deeper than a typical VM hypervisor.

You cannot claim with a straight face that Virtualbox is easier to use.

It's deeper but let's not overblow it.

I think the two fairly deep integrations are window's ability to navigate WSL's filesystem and wslg's fairly good ability to serve up guis.

The filesystem navigation is something that AFAIK can't easily be replicated. wslg, however, is something that other VMs have and can do. It's a bit of a pain, but doable.

What makes WSL nice is the fact that it feels pretty close to being a native terminal that can launch native application.

I do wish that WSL1 was taken further. My biggest grip with WSL is the fact that it is a VM and thus takes a large memory footprint. It'd be nice if the WSL1 approach panned out and we instead had a nice clean compatibility wrapper over winapi for linux applications.

> The integration between Windows and the WSL VM is far deeper than a typical VM hypervisor.

Sure, but I never claimed otherwise.

> You cannot claim with a straight face that Virtualbox is easier to use.

I also didn't claim that. I wasn't comparing WSL to other virtualization solutions.

WSL2 is cool. Linux doesn't have a tool like WSL2 that manages Linux virtual machines.

The catch 22 is that it doesn't need one. If you want to drop a shell in a virtual environment Linux can do that six ways through Sunday with no hardware VM in sight using the myriad of namespacing technologies available.

So while you don't have WSL2 on Linux, you don't need it. If you just want a ubuntu2204 shell or something, and you want it to magically work, you don't need a huge thing with tons of integration like WSL2. A standalone program can provide all of the functionality.

I have a feeling people might actually be legitimately skeptical. Let me prove this out. I am on NixOS, on a machine that does not have distrobox. It's not even installed, and I don't really have to install it since it's just a simple standalone program. I will do:

    $ nix run nixpkgs#distrobox enter

Here's what happened:

    $ nix run nixpkgs#distrobox enter
    Error: no such container my-distrobox
    Create it now, out of image registry.fedoraproject.org/fedora-toolbox:latest? [Y/n]: Y
    Creating the container my-distrobox
    Trying to pull registry.fedoraproject.org/fedora-toolbox:latest...
    ...
    0f3de909e96d48bd294d138b1a525a6a22621f38cb775a991974313eda1a4119
    Creating 'my-distrobox' using image registry.fedoraproject.org/fedora-toolbox:latest [ OK ]
    Distrobox 'my-distrobox' successfully created.
    To enter, run:

    distrobox enter my-distrobox

    Starting container...                    [ OK ]
    Installing basic packages...             [ OK ]
    Setting up devpts mounts...              [ OK ]
    Setting up read-only mounts...           [ OK ]
    Setting up read-write mounts...          [ OK ]
    Setting up host's sockets integration... [ OK ]
    Integrating host's themes, icons, fonts... [ OK ]
    Setting up distrobox profile...          [ OK ]
    Setting up sudo...                       [ OK ]
    Setting up user groups...                [ OK ]
    Setting up user's group list...          [ OK ]
    Setting up existing user...              [ OK ]
    Ensuring user's access...                [ OK ]

    Container Setup Complete!
    [john@my-distrobox]~% sudo yum install glxgears
    ...
    Complete!
    [john@my-distrobox]~% glxgears
    Running synchronized to the vertical refresh.  The framerate should be
    approximately the same as the monitor refresh rate.
    302 frames in 5.0 seconds = 60.261 FPS
    ^C

No steps omitted. I can install software, including desktop software, including things that need hardware acceleration (yep, even on NixOS where everything is weird) and just run them. There's nothing to configure at all.

That's just Fedora. WSL can run a lot of distros, including Ubuntu. Of course, you can do the same thing with Distrobox. Is it hard? Let's find out by using Ubuntu 22.04 instead, with console output omitted:

   $ distrobox create --image ubuntu:22.04
   ...
   $ distrobox enter ubuntu-22-04
   ...
   $ sudo apt install openarena
   ...
   $ /usr/games/openarena

To be completely, 100% fair: running an old version of Ubuntu like this does actually have one downside: it triggers OpenGL software rendering for me, because the OpenGL drivers in Ubuntu 22.04 are too old to support my relatively new RX 9070 XT. You'd need to install or copy in newer drivers to make it work. There are in fact ways to do that (Ubuntu has no shortage of repos just for getting more up-to-date drivers and they work inside Distrobox pretty much the same way they work in real hardware.) Amusingly, this problem doesn't impact NVIDIA since you can just tell distrobox to copy in the NVIDIA driver verbatim with the --nvidia flag. (One of the few major points in favor of proprietary drivers, I suppose.)

On the other hand, even trying pretty hard (and using special drivers) I could never get hardware acceleration for OpenGL working inside of WSL2, so it could be worse.

That aside, everything works. More complex applications (e.g. file browsers, Krita, Blender) work just fine and you get your normal home folder mapped in just like you'd expect.

Yes, yes I can. Also does most of everything. WSL has severe issues with hardware translation.

> I get that WSL is revolutionary for Windows users

It is... I'm working these days on bringing a legacy windows only application to the 21st century.

We are throwing a WSL container behind it and relying on the huge ecosystem of server software available for Linux to add functionality.

Yes that stuff could run directly on windows, but you'd be a lot more limited in what's supported. Even for some restricted values of supported. And you'd have to reinvent the wheel for a few parts.

And if they think that this version of Linux "isn't janky" but regular Linux is, than idk what to say.

With WSL you can use “Linux the good parts” (command line tools, efficient-enough paradigms for fork() servers) and completely avoid X Windows, the Wayland death spiral, 100 revisions of Gnome and KDE that not so much reinvent the wheel but instead show us why the wheel is not square or triangular…

This is very much YMMV thing. There is no objectively best platform. There are different users and requirements.

I’ve been a software developer for 20 years and in _my_ opinion Windows is the best platform for professional software development. I only drop of to linux when need some of the excellent posix tools but my whole work ergonomy is based on Windows shortcuts and Visual Studio.

I’ve been forced to use Mac for the past 1.5y but would prefer not to.

Why would Windows be superior for me? Because that’s where the users are (for the work stuff I did before this latest gig). I started in real time graphics and then spent over a decade in CAD for AEC (developing components for various offerings including SketchUp). The most critical thing for the stuff I did was the need to develop on the same platform as users run the software - C++ is only theoretically platform independent.

Windows API:s are shit for sure for the most part.

But still, from this pov, WSL was and will be the best Linux for me as well.

YMMV.

I fully agree with you - "YMMV" is the one true take. Visual Studio has never been particularly attractive to me, my whole workflow is filled with POSIX tools, and my code mostly runs on Docker and Linux servers. Windows is just another thing to worry about for me, be it having to deal with the subtle quirks of WSL not running on raw metal or having to deal with running UNIX-first tooling (or finding alternatives) on Windows. If it wasn't for our work provided machines being Windows by default, and at home, being into VR gaming and audio production (mostly commercial plugins), I'd completely ditch Windows in a heartbeat.

It's a VM plus some usability automation. Can't ignore the usability benefits.

Just FYI, you may also enjoy systemd-machine. It's essentially the same thing as toolbx but it handles the system bus much more sanely, and you can see everything running inside the guest from the host's systemctl.

If Windows provided easier access to hardware, especially USB, from WSL it would be nice. In fact, if WSL enumerated devices and dealt with them as native Linux does, even better.

Windows has many useful software that is not available on Linux.

So, for me Windows + WSL is more productive than just using Linux. The UI is still better on Windows(basic utilities like File Explorer and Config Management is better on Windows). No Remoting Software beats RDP. When I remote to a Windows workstation through RDP, I can't tell the difference. VNC is always janky. Of course there is Word/Excel/Illustrator which is simply not available on Linux

File Explorer is better on Windows? How? I tried Windows 11 for the first time a month ago and it takes several seconds for file explorer to open, it's asynchronously loading like 3 different UI frameworks as random elements pop in with no consistency, there's two different rightclick menus because they couldn't figure out how to make the new one have all the functionality of the old one so they decided to just keep the old one behind "Show More Options", and it's constantly pushing OneDrive in your face. I'm offended that this is what they thought is good enough to ship to a billion users.

> No Remoting Software beats RDP. When I remote to a Windows workstation through RDP, I can't tell the difference. VNC is always janky

Any recent distro running Gnome or KDE has built-in support for connecting and hosting an RDP session. This used to be a pain point, you don't need to use VNC anymore.

It's actually worse on windows since you need to pony up for a pro license to get RDP hosting support...

> The UI is still better on Windows(basic utilities like File Explorer and Config Management is better on Windows).

5 years ago, we would be comparing old GNOME 3 or KDE Plasma 5 on X11 and Windows 10. I would be forced to agree. The Windows UI was better in many ways at that point.

Today we have KDE Plasma 6.3 on Wayland and Windows 11. This is an entirely different ball game. It's hard to explain. Wayland feels like it has taken an eternity to lift off, like well over a decade, but now things change dramatically on the scale of months. A few months ago HDR basically didn't work anywhere. Right now it's right in front of me and it works great. You can configure color profiles, SDR applications don't break ever, and you even get emulated brightness. Display scaling? Multiple monitors with different scale factors? What about one monitor at 150% and another at 175% scale factor? What about seamlessly dragging windows between displays with different scale factors? Yes, Yes, Yes, and Yes. No `xrandr` commands. You configure it in the GUI. I am dead serious.

File Explorer? That's the application that has two context menus, right? I think at this point Windows users might actually be better off installing KDE's Dolphin file manager in Windows for the sake of their own productivity. If I had the option to use Windows File Explorer on KDE I would impolitely decline. I have not encountered any advertising built into my file explorer. I do not have an annoying OneDrive item in the menu on the left. I have a file tree, a list of Places, and some remote file shares. When I right click it does not freeze, instead it tends to show the context menu right away. And no, I'm not impressed by Tabs and Dark Mode, because we've had that on Linux file managers for so long that some people reading this were probably born after it was already supported.

Windows still has the edge in some areas, but it just isn't what it used to be. The Linux UI is no longer a toy.

> When I remote to a Windows workstation through RDP, I can't tell the difference. VNC is always janky.

I don't really blame you if you don't believe me, but I, just now, went into System Settings, went to the Remote Desktop setting, and clicked a toggle box, at which point an RDP server spawned. Yes, RDP, not VNC, not something else. I just logged into it using Reminna.

Not everything on Linux is seamless and simple like this, but in this case it really is. I'm not omitting a bunch of confusing troubleshooting steps here, you really can do this on a modern Linux setup, with your mouse cursor. Only one hand required.

> Of course there is Word/Excel/Illustrator which is simply not available on Linux

True, but if you want to use Linux and you're held back by needing some specific software, maybe it's not the end of the world. You have many options today. You can install VirtualBox and run your spreadsheets in there. You can use Office 365 in a browser. You can run Crossover[1] and emulate it. You can use an office alternative, like possibly WPS Office. You can dual boot. You can go the crazy route and set up a KVM GPU passthrough virtual machine, for actually native performance without needing to reboot.

The point I'm making here is not "Look, Linux is better now! Everyone go use it and get disappointed ASAP!" If you are happy with Windows, there's literally no point in going and setting yourself up for disappointment. Most people who use Linux do so because they are very much not happy with Windows. I'm sure you can tell that I am not. However, in trying to temper the unending optimism of Linux nerds, sometimes people go too far the other way and represent Linux as being in far worse of a state than it actually is. It really isn't that bad.

The worst thing about modern Linux is, IMO, getting it to work well on your hardware. Once you have that part figured out, I think modern Linux is a pretty good experience, and I highly recommend people give it a shot if they're curious. I think Bazzite is a really nice distro to throw on a random spare computer just to see what modern Linux is actually capable of. It's not the absolute most cutting edge, but it gives you a nice blend of fairly up-to-date software and a fairly modern RPM ostree base system for better stability and robustness, and it's pretty user-friendly. And if you don't like it, you can easily get a full refund!

[1]: https://www.codeweavers.com/compatibility/crossover/microsof...

The last time I deployed Linux servers on bare metal was about 2010.

Apparently Linux VMs on other people's computers is very much appreciated.

I definitely prefer working in Linux.

But having Windows tightly integrated when needed is nice.

If only I could run replace the Windows shell with a Linux DE...

Is it a VM? It seems to be much faster than most VMs I've used.

Literally built on top of MS's Hyper-V.

IDK how many VMs you've used, but there has been a lot of work specifically with x86 to make VMs nearly as fast as native. If you interact with cloud services everything you do is likely on a VM.

It's handy if you have other services that are Windows-based, though. And, being a VM, it's fairly convenient to have multiple versions and to back up.

So, how you run Windows on Linux like WSL does?

Methods I know are using qemu/Wine/proxmox/VirtualBox.

> You know what's even more convenient than a VM? Not needing a VM and still having the exact same functionality.

Exactly.

I heard 2025 was the year of Linux on the desktop!

> You know what's even more convenient than a VM? Not needing a VM and still having the exact same functionality

I mean this is basically heresy now.

most code is virtualised, or sandboxed, or in a VM, or a docker container, or several of the above at the same time.

The important bit though is that Docker containers are not VMs or sandboxes, they're "just" a combination of technologies that give you an isolated userland using mostly Linux namespaces. If you're running a Linux host you already have namespaces, so you can just use them directly. Distrobox gives you basically the same sort of experience as WSL2 except it doesn't have any of the weird parts of running a VM because it's not VMs.

Your comment that you can do Linux things on Linux missed the point entirely.

Where is the reverse WSL on Linux, where Windows is deeply embedded and you have all the Windows features in your hands?

You can use Wine/Crosseover, which is cool, but even now the number of software products it supports is tiny. Steam has a lot of games.

You can run a virtual machine with Windows on it. That is identical to what you can do on Windows with Linux.

WSL2-> is a virtual machine with unique tooling around it that makes it easier to use and integrates well with Windows.

Windows supports Linux because the latter is open source, it's a lot easier than the reverse.

Linux, on the other hand, barely supports Windows because the latter is closed, and not just closed, windows issues component updates which specifically check if they run in wine and stop running, being actively hostile to a potential Linux host.

The two are not equivalent, nobody in the Linux kernel team is actively sabotaging WSL, whereas Microsoft is actively sabotaging wine.

I preferred WSL to running linux directly even though I had no need for any windows only software. Not having to spend time configuring my computer to make basic things work like suspend/wake on lid down/up, battery life, hardware acceleration for video playback on the browser, display scaling on external monitor and so on was reason enough.

> You can use Wine/Crosseover, which is cool, but even now the number of software products it supports is tiny. Steam has a lot of games.

This isn't really the case, and hasn't been for some years now, especially since Valve started investing heavily in Wine. The quality of Wine these days is absolutely stunning, to the point that some software runs better under Wine than it does on Win11. Then there's the breadth of support which has has moved the experience from there being a slight chance of something running on Wine, to now it being surprising when something doesn't.

> Where is the reverse WSL on Linux, where Windows is deeply embedded and you have all the Windows features in your hands?

https://github.com/Fmstrat/winapps

Enjoy.

> WSL is more powerful than Linux

This is the kind of statement that makes you pay the karma tax. WSL is great, I use it on a day to day basis. I also use Linux on a day to day basis. And as great as WSL is, for running Linux software on supported hardware, Linux beats WSL hands down. And I mean, of course it does, do you expect a VM to beat native? In the same way that Windows software runs better on Windows. (with a few exceptions on both sides).

Compared to Linux, WSL I/O is slow, graphics is slow and a bit janky, I sometimes get crashes, memory management is suboptimal, networking has some quirks, etc... These problems are typical of VMs as it is hard for the host and guest OS to coordinate resource use. If you have an overpowered computer with plenty of RAM, and are mostly just using the command line, and don't do anything unusual with your network, then sure it may be "better" than Linux. But the truth is that it really depends on your situation.

Well, WSL is Linux. It's really just a VM of it (since WSL2, WSL1 was actually running on the windows kernel which was pretty cool).

The big drawback to WSL to me is the slow filesystem access because NTFS sucks. And having to deal with Windows in the first place.

Ps I wouldn't worry about your karma. It's just a number :P

NTFS is not the problem.

The problem is Windows IO filters and whatnot, Microsoft Defender trying to lazily intercept every file operation, and if you're crossing between windows and Linux land, possibly 9pfs network shares.

WSL2's own disk is just a VM image and fairly fast - you're just accessing a single file with some special optimizations. Usually far, far more responsive than anything done by windows itself. Don't do your work in your network-shared windows home folder.

>The problem is Windows IO filters

Not the biggest issue of them, 'find' and 'git status' on WSL2 in a big project is still >100 times slower on windows dev drive which avoids those filters than it is with WSL 1 on dev drive.

WSL 1 on regular ntfs with defender disabled is about 4x slower than WSL1 on dev drive, so that stuff does cause some of it, but WSL2 feels hopelessly slow. And wsl 2 can't share memory as well or take as much advantage of the filesystem cache (doubling it if you use the windows drive in both places I think, unless the network drive representation of it doesn't get cached on the WSL2 drive.

WSL2, in my testing, is orders of magnitude faster at file heavy operations than anything outside WSL, dev drive or not. We have an R&D department that's using WSL2 and jumping through hurdles of forwarding hardware because it's night and day compared to trying under windows on the same machine. It provided other benefits too, but the sheer performance was the main selling point.

WSL2 does not take less advantage of filesystem caches. Linux's block cache is perfectly capable. HyperV is a semi-serious hypervisor, so it should be using a direct I/O abstraction for writing to the disk image. Memory is also balloning, and can dynamically grow and shrink depending on memory pressure.

Linux VM's is something Microsoft has poured a lot of money into optimizing as that's what the vast majority of Azure is. Cramming more out of a single machine, and therefore more things into a single machine, directly correlates with profits, so that's a heavy investment.

I wonder why you're seeing different results. I have no experience with WSL1, and looking into a proprietary legacy solution with known issues and limited features would be a purely academic exercise that I'm not sure is worth it.

(I personally don't use Windows, but I work with departments whose parent companies enforce it on their networks,

> Linux's block cache is perfectly capable. HyperV is a semi-serious hypervisor, so it should be using a direct I/O abstraction for writing to the disk image.

Files on the WSL2 disk image work great. They're complaining about accessing files that aren't on the disk image, where everything is relayed over a 9P network filesystem and not a block device. That's the part that gets really slow in WSL2, much slower than WSL1's nearly-native access.

> Memory is also balloning, and can dynamically grow and shrink depending on memory pressure.

In my experience this works pretty badly.

> a proprietary legacy solution with known issues and limited features

Well at least at the launch of WSL2 they said WSL1 wasn't legacy, I'm not sure if that has changed.

But either way you're using a highly proprietary system, and both WSL1 and WSL2 have significant known issues and limited features, neither one clearly better than the other.

> WSL2 does not take less advantage of filesystem caches.

My understanding is when you access files on the windows drive, the linuxvm in WSL2 caches it in its own memory, and the windows side caches it in its: now you have double the memory usage on disk cache where files are active on both, taking much less advantage of caches than if you had used WSL1 where windows serves as the sole cache for windows drives.

I'm only comparing working on windows filesystems that can be accessed by both. My use case is developing on large windows game projects, where the game needs the files fast when running, and WSL needs the files fast when searching code, using git, etc. WSL1 was usable on plain NTFS, and now much closer to ext4 with dev drive NTFS. WSL2 I couldn't make fast.

You could potentially have the windows files on a network drive on the WSL2 side living in native ext4, but with that you get the double filesystem caching issue, and you might slow a game editor launch on the windows side by way too much, your files are inaccessible during upgrades and you have to always have RAM dedicated to WSL2 running to be able to read your files. MS store versions of WSL2 will even auto upgrade while running and randomly make that drive unavailable.

Running WSL2 on Dev Drive means that you're effectively doing network I/O (to localhost); of course it's slow. It's also very pointless since your WSL2 FS is already a separate VHD.

Not pointless if you are working on a windows project but using unix tools to search code, do commits, etc. WSL2 just isn't usable for it in large projects. git status can take 5 minutes on unreal engine.

If you just need the stock Unix command line tools, MSYS2 will give you them at native speed, no VM needed, no funky path mappings etc.

WSL is for when you actually need it to be Linux.

I don't need just stock commandline tools, I use Claude Code etc. Maybe it can work in msys2 (it uses node) but it works well in WSL1. Also do a good bit with Python where some dependencies haven't worked in msys2.

I do windows and android gamedev and the server side is Linux.

However, WSL1 is pretty much abandoned and lots of newer distros have moved to forcing systemd or container based stuff that doesn't work in it, and some elf binaries no longer work in it (including newer node.js versions).

I use WSL2 as well when I really need it to be Linux and am doing webdev or something.

I use msys2 for a few things, usually compiling windows dependencies with the gnu toolchain that need GCC's stuff like inline assembly or computed goto (codecs etc.). Maybe possible from wsl1 too but they usually have full build and dependency instructions for msys2.

I use it, I am required to use Windows, and it’s a huge improvement over doing Data Science on native Windows, but the terrible filesystem access ruins what otherwise would be a seamless experience.

It’s fine for running small models but when you get to large training sets that don’t fit in RAM it becomes miserable.

There is a line where the convenience of training or developing locally gives way to a larger on demand cloud VM, but on WSL the line is much closer.

Slow IO is why I still use wsl1.

This. WSL was SO much more interesting in v1 times.

I liked the networking in WSL1 more too

Corporate networking is why I still use WSL1 (I didn’t spend enough time to check why it doesn’t with WSL2, zScaler could be the culprit maybe).

However it’s not perfect, for example I hit this bug when trying to run node a few days ago https://github.com/microsoft/WSL/issues/8219#issuecomment-10... and I don’t think they’re fixing bugs in WSL1 anymore

still use WSL1 also because VMWare runs so dreadfully slow with any kind of Hyper-V enabled - if so, VMWare must also use it, so you get a Type-2 running under a Type-1 the lag is untennable lag and performance.

>The big drawback to WSL to me is the slow filesystem access because NTFS sucks

Thats if you are going from VM/host. If you use the allocated space for VM, its pretty fast.

Is it really a NTFS issue ?

The culprit would be the plan9 bits (think of smb or nfs but .. wilder ? why are they using 9P again ?)

I'm guessing they use plan9 because distros already ship support for it, and it's super simple compared to NFS? It doesn't seem like CIFS/NFS would be any faster, and they introduce a lot more complexity.

Where are you experiencing filesystem slowness? I've been using WSL in some advanced configurations (building Win32 apps by cross-compiling from Linux CLANG and dropping the .exe into a Windows folder, copying large files from Linux->Windows and vice versa, automating Linux with .BAT files, etc.) and I haven't seen this slowness at all.

> NTFS sucks

Watch https://www.youtube.com/watch?v=qbKGw8MQ0i8 please.

While I can see the subtle distinction you're trying to draw people's attention to (NTFS is not the problem, filesystem operations generally on Windows are the problem) I have to say it seems like a distinction without a difference in real terms. They made a range of changes that seem to produce more complicated code everywhere because the overhead of various filesystem tasks are substantially higher on this OS vs every other OS.

But in the end they had to get the OS vendor to bless their process name anyway, just so the OS would stop doing things that tank the performance for everybody else doing something similar but who haven't opened a direct line up with the OS vendor and got their process name on a list.

This seems like a pain point for the vendor to fix, rather than everybody shipping software to their OS

I find it to be incredibly janky. Pretty much every every time my computer sleeps (so every morning, at least) I have to restart it because somehow the VM-host networking gets screwed up and VS code connections into the VM stop working. You also can't just put things in your Windows User directory because the filesystem driver is so slow that git commands will take multiple seconds, so now you have two home directories to keep track of. There were also some extremely arcane things I had to fix when setting it up involving host DNS and VPN adapter priority not getting propagated into the VM so networking was completely broken. IIRC time would also not match the host after a sleep and get extremely far out of sync, though I haven't run into that for a while since now I have to reboot Windows constantly anyway.

I don't have a need to run multiple OSes though. All of my tools are Linux based, and in companies that don't let people run Linux, the actual tools of the trade are almost all in a Linux VM because it's the only reasonable way to use them, and everything else is cross-platform. The outer OS just creates needless issues so that you now need to be a power user with two operating systems and their weird interactions.

> somehow the VM-host networking gets screwed up

> extremely arcane things I had to fix when setting it up involving host DNS and VPN adapter priority not getting propagated into the VM so networking was completely broken

Are you sure you set up the VPN properly? Messing around with Linux configs is a good way to end up with "somehow" bugs like that.

I don't know how it's set up. That's kind of my point though. I have to now be an expert in Linux and Windows to debug this stuff, which is a waste of my time as someone who's job it is to develop (server, i.e. Linux) software. I had exactly zero issues when I was using Fedora. At one point my company made all of the Linux users move off (we do now have an IT-supported Linux image, but I haven't found the time to re-set up my laptop and don't fully trust that it will work without a bunch of trouble/IT back-and-forth because they also made Windows users start using passkeys), and since then I've seen way more issues with Windows than Linux (e.g. one day my start menu just stopped reacting to me clicking on programs), in addition to things like ads in the lock screen and popups for some XBox pass thing that I had to turn off, which is just insane in a "professional" OS. A lot of days I end up having to hold down the power button to reboot because it just locks up entirely.

OSX was a bit janky with docker filesystem slowness, homebrew being the generally recommended package manager despite being awful (why do I sometimes tap a cask and sometimes pour a bottle? Don't tell me; I don't care. Just make it be "install". Also, don't take "install" as a cue to go update all of my other programs with incompatible versions without asking), annoying 1+ second animations that you can't turn off that make it so the only reasonable way to use your computer is to never maximize a window (with no tiling support of course), and completely broken external monitor support (text is completely illegible IIRC), but Windows takes jank to another level.

By contrast, I never encounter the issues people complain about on Linux. Bluetooth works fine. Wifi works fine. nVidia GPUs and games work fine. Containers are easy to use because they're natively part of the OS. I prefer Linux exactly because I stopped enjoying "tinkering" with my computer like 10 years ago, and I want it to just quietly work without drawing attention to itself (and because Windows 8 and the flat themes that followed were hideous and I was never going to downgrade to that from Windows 7).

Thats odd. I have none of these problems. Sleep doesnt interrupt the VM. And I regularly use the git CLI through WSL on projects living within windows user directories. Both work fine.

FWIW, you can run a VPN (e.g. tailscale) in WSL2. I have WSL2 start up on boot and I can remotely ssh to WSL2 without logging into Windows at all.

I also have tailscale running on Windows itself and they don't conflict.

I think you might want to give more context.

I use linux. I don't need WSL at all. Not at work nor at home.

So you praise WSL because you use Windows as your main system? Than yes its great. It definitly makes the Windows experience a lot better.

OpenSSH for Windows was also a game changer. Honestly, i have no clue why Microsoft needed so long for that.

Openssh should have been a game changer but they made a classic openssh porting bug (not reading all bytes from the channel on close) and have now been sat on the fix in “prerelease” for years. I prodded the VP over the group about the issue and they repeatedly made excuses about how the team is too small and getting updates over to the windows team is too hard. That was multiple windows releases ago. Over on GitHub if you look up git receive pack errors being frequent clone problems for windows users you’ll find constant reports ever since the git distribution stopped using its own ssh. I know a bunch of good people at Microsoft, but this leadership is incapable of operating in a user centric manner and shouldn’t be trusted with embedded OSS forks.

I'm a simple man, if I open the shell and `ssh foo@bar.com` doesn't work, I don't use that computer. Idk if Windows has fixed that yet or why it's so hard for them. Also couldn't even find the shell on a Chromebook.

putty is longer necessary? That would be a wild upgrade in usability for the work laptop, shall go try it

openssh has been an optional windows component for... almost a decade now? including the server, so you can ssh into powershell as easily as into any unix-like. (last time I set it up there was some fiddling with file permissions required for key auth to work, but it does work.)

I dislike using putty, I use the ssh client from WSL. Just feels .. better. And bash/fish history helps.

https://xkcd.com/963/

On the other hand sometimes the GUI on WSL decides to break and you have to restart the whole thing.

Aged like fine milk

Running a Linux VM on Windows is nicer than just booting into Linux? That's quite a take. Windows is so user-hostile these days that I feel bad for those who have to deal with it. Calling it delightful must be symptomatic of some sort of Stockholm syndrome.

> symptomatic of some sort of Stockholm syndrome

I have since moved to macbooks for the hardware, but until not too long ago WSL was my linux "distro" of choice because I didn't want to spend time configuring my computer to make basic things work like suspend/wake on lid down/up, battery life, hardware acceleration for video playback on the browser, display scaling on external monitor and so on.

> Running a Linux VM on Windows is nicer than just booting into Linux

Indeed, it does. Having stable system and not dealing with Linux on Desktop, clear tradoffs (like "just add another 16gb RAM stick in laptop/desktop and you are golden") is great for peace of mind.

The average uptimes on my laptops (note for plural) is ~3 weeks, until next Windows Update to be applied. I don't have nostalgia on the days of using Linux on desktop (~2003 student times, ~2008 giving it one more try, ~2015 as required by dayjob)

Of course it adds up that I can tell people around me (who are not tech guys often, but smart enough to know basic concepts and be able to run bash scripts provided to them) - "yep, machine with 32GB+ of RAM will work fine, choose any you like" - and it works.

If for some reason I could never use a MacBook again, it wouldn't be easy to decide between Windows or Linux as the host OS on a laptop. Do I want something that's intentionally user-hostile or something that's unintentionally broken a lot?

I'd at least try Linux cause I abhor Microsoft, but idk if it'd work out.

Maybe it is both-sidesism but the motd you get by default on Ubuntu these days is as bad as any OS. (“Ubuntu Advantage” sounds about as good as https://prospect.org/health/2024-01-12-great-medicare-advant...)

At least the nags in Windows look like modern web-based UI (so far that ‘use Electron’ seems to be the post-Win 8 answer to ‘how to make Windows apps’) in contrast to MacOS which drove my wife crazy with nag dialogs that look like a 1999 refresh of what modal dialogs looked like on the classic Mac in 1984.

My acid test for WSL2 was to install the Linux version of Google Chrome in it, and then play Youtube videos fullscreen with that. It worked. Somehow WSL1 was the more impressive hack but how can you argue with what works? WSL2 works fine.

Also 1980s style X11 widgets on the Windows desktop in their own windows? Cool.

I have to say too, though, once you get the hang of the way an EFI system boots, it's really good for dual boot. I let the Linux installer mount the undersized existing one as /boot/orig_efi and made a new, bigger EFI system partition. Not only was the UEFI on that particular laptop fine with it, scanning both EFI system partitions for bootable stuff, but also, grub2 installed in the new one automatically included the Windows boot in the old one as a boot option.

Cool because nothing about how Windows boots is intercepted; you can just nuke the new partitions (or overwrite them with a new Linux installation). I still prefer a native Linux boot with "just in case" Windows option to WSL.

But not having to dual boot and just get both worlds at the same time definitely beats having to dual boot.

I don't think people are using WSL to avoid problems with dual booting. Dual-booting has become about as simple as it can be, thanks to UEFI, but it's still not exactly fun to have to close all of your open apps to switch to another OS to run just one app.

You get much nicer window decorations if you use the wayland support instead of X11.

> You get much nicer window decorations if you use the wayland support instead of X11.

Wayland supports window managers ?

Step it up a notch and see if Netflix works w/ its DRM.

Forced to work on Windows for ++nth job, I was looking forward to WSL. Indeed, while it worked, it was magic. Sadly, I have had no end of bizarre bugs. The latest one almost crashed my whole desktop - as far as I can piece together, something crashed, leading to a core dump the size of my desktops entire memory - half the machine's RAM. This in turn put WSL in a weird state - it would neither run, not be uninstallable. Googling found bug reports with similar experiences, no responses from Microsoft and magic incantation that maybe worked for some people - but not for me.

It might be due to my corpo's particular setup etc. but for me 95% of the value of WSL would be the ability to run it on "corporate" Windows boxes. Alas.

I'm sure that feature is important for whatever works you're doing, but that's a feature I've _never_ desired, and WSL is missing plenty of features that are important for my work.

Hardware performance counters basically do not work in WSL2, which among other issues, makes it extremely difficult to use rr. https://github.com/rr-debugger/rr/issues/2506#issuecomment-2... Some people say they got it working, but I and many other users encounter esoteric blockers.

The Dozen driver is never at feature parity with native Linux Vulkan drivers, and that's always going to be the case.

By default, WSL security mitigations cause GCC trampolines to just not work, which partly motivated the opt-in alternative implementations of trampolines last year. https://gcc.gnu.org/git/?p=gcc.git;a=commit;h=28d8c680aaea46...

gWSL is also a terrible X11 server that makes many very basic window management configurations impossible, and while I prefer VcXsrv, it has its own different terrible issues.

I can imagine that WSL2 looks attractive if all you want to do is run command line apps in multiple isolated environments, but it is miserable for anything graphical or interactive.

> I can imagine that WSL2 looks attractive if all you want to do is run command line apps in multiple isolated environments, but it is miserable for anything graphical or interactive.

Indeed, that's my case - using CLI mostly for ssh/curls/ansible/vim over ansible and Puppet, so on.

For GUI part, Windows is chosen and shines for me.

I think it really depends on what you do and whether the Linux side of it has hard dependencies on system packages. Personally, at work I much prefer working directly on my Linux workstation, and at home have even switched to using Linux for my gaming desktop. I really don't like the direction Windows has been trending for the past few years, and with the specter of a forced Windows 11 upgrade on the horizon I decided it's time to go all in. My system runs better and I can still play all my games. The jankiest thing I do is I have a mingw toolchain so I can compile some game mods into Windows DLLs to be loaded by Wine, but even that ended up being pretty seamless. Just install the toolchain and the project just compiled.

I don't understand. Docker/podman/distrobox/lxc all allow you to do the exact same thing without the virtual machine overhead. I think the real win of WSL is that its a best of all worlds. You get to use Windows with access to every game ever made plus all of the proprietary apps everyone needs to use, with all of the upside of having a full and complete linux command line experience.

You get all of Windows telemetry, vulnerabilities and backdoors, the always fun game of spot the new Advertising opportunity, AI “copilot” spyware I mean feature, updates that reset your machine at will, a terrible UAC model that encourages “just click OK already!”, and dependence on a company that has gone out of their way to prove how much of an unstoppable behemoth they are; and best of all you get to pay for the privileges above.

I know… every year is the year of the Linux desktop… but seriously the AI spyware included was enough to get me gone for good.

It's hard to pick the Windows feature I hate the most, but floating around at the top is Defender. It can't be disabled, at least not easily, and it demolishes IO performance. And Windows update takes the computer hostage, and takes ages to do anything giving no feedback in the process, meanwhile APT can update to a new major version in like 5-10 minutes.

You can setup local and limited user accounts under Windows. Many applications including every development tool out there doesn't need any admin permissions.

Spyware and adware is a government policy / regulation problem. Thanks to GDPR and DMA, using Windows in EU is significantly better experience (try setting a Windows desktop with an EU image). You can remove almost all of the apps including Edge and Copilot. There are no ads in the UI. Neither in Explorer nor in Start menu.

Yes, you get Windows telemetry which enabled fixing bugs without a bug report, you get minimal ads in the start menu (if you're playing "spot the new advertising opportunity" I found it. It's in the start menu. You can stop playing now), AI "copilot" which isn't spyware just because you think it is, updates that ASK you nicely multiple times to update (I don't want to be ableist, if you suffer from a Christopher Nolan Memento-like disability where you don't remember the warnings, you might think it's "resetting at will", but I assure you, it isn't), a great UAC model that's a lot better than "just type your root password into this terminal already, and just hope the binary wasn't hijacked in some way to keylog you, because unlike UAC, there is no visual evidence that you're not getting hacked", and dependence on a company that SV_BubbleTime thinks "has gone out of their way to prove how much of an unstoppable behemoth they are" with no evidence or clarity so they must just be making FUD, and best of all the OS costs so little you can pay it in 8 hours of working as a software developer.

with the virtual machine overhead.

Because it's easier to set up a local dev environment in WSL than in any of those.

How is it easier to setup a linux dev environment in WSL than in https://containertoolbx.org/ or https://distrobox.it/ or just in Linux directly?

Gnome (a linux desktop environment) ships a "Boxes" app [0] that is very impressive. You can, with a few clicks, install one of a huge number of Linux distros in an auto-provisioned VM, enable hardware passthrough for USB devices and host 3D acceleration, and manage files with drag-and-drop from the host system. I also use it for Windows and MacOS VMs (don't tell Apple), but you need to provide your own images.

[0]: https://apps.gnome.org/Boxes/

> WSL is more powerful than Linux ...

Are you a Windows user who is happy to have a good way to run Linux on Windows, or are you a Linux user trying to convince other Linux user that instead of using Linux, they should use Linux in a VM running on Windows?

I am a longtime Linux user, and I can't see a reason in the universe why I would want to access my Linux through a VM on Windows. That seems absolutely insane.

Look I get it. I’m forced to use Windows at work and I thank the lord WSL is a thing. But I would switch to Linux base in a heartbeat if I could. WSL is jank as fuck compared to just using Linux.

I will also die on this hill - NixOS on WSL + Windows + komorebi[1] for tiling window management is peak productivity for me.

[1]: https://github.com/LGUG2Z/komorebi

Why not a Linux distro with i3wm, instead? What could possibly hold you back from upgrading?

I've yet to find anything comparable feature-wise on Linux - and they all come with the huge downside of having to roll your own cohesive settings widget ecosystem for basic everyday things like WiFi and Bluetooth connectivity. I run Cosmic Epoch on my old Macbook which is better, but again, feature-wise, it's just not comparable for serious work.

I just run NixOS, but that feels like a respectable answer.

> WSL is more powerful than Linux because of how easy it is to run multiple OS on the same computer simultaneously.

I'd venture to say this depends on which OS you're more comfortable with. I'm more comfortable with Linux, so I'd say it's easier/better/less janky to use Linux as a host OS.

> Like if one project has a dependency on Ubuntu22 and another is easier with Ubuntu24. You don't have to stress "do I update my OS?"

Once you're a developer who's been burned by this enough times, you do this with containers or dedicated dev VMs. You do not develop on your host OS and stay sane.

I think it depends a lot on what you're trying to do. I found that anything GPU-related was a nightmare of drivers and configuration which was a show-stopper for me. Now I just run arch/kde and that all works fine out of the box

Well, I'd still rather just use linux, but I take your meaning.

Me too. Particularly after having to do Docker things a few years ago, destroying my productivity due to file system speed.

However, for those of us that went Linux many years ago, and like our free open source, in 2025, is it better to go back to the dark side, to run Windows and have things like a LAMP stack and terminals run with WSL?

I don't play games or run Adobe products, I use Google Docs and I don't need lots of different Linux kernels. Hence, is it better to run Linux in Windows now? Genuinely asking.

As someone who occasionally does use WSL, I definitely think it's not better no. But I'm still biased, because I know a lot more about using linux than I do about using windows, and WSL is still windows.

for me,

> is it better to run Linux in Windows now? Genuinely asking.

definitely is. Servicing takes ~ 1 minute per month to click on "yeah, let's apply those updates and reboot". Peace of mind with no worrying on external hardware won't work or monitor will have issues or laptop won't sleep or during the call battery will discharge faster due to lack of hardware acceleration or noise cancellation not working or ...

wsl2 is linux

I agree. Back in the day (10+ years ago), I used to argue with people about why I ran VMs instead of just partitioning the disk and booting up the OS I needed.

XAMPP did not work out of the box with me on Windows (skill issue on my part, I know), so my preferred setup was to run a Ubuntu Server VM (LAMP stack) and then develop whatever I had on a Windows IDE.

I could have done that under full Linux, I just did not want that. Then Vagrant came into existence, which I'd say was for my use case (but never came around to adopt it).

I'm really happy with my WSL2 setup. I stopped using VMware Workstation when WSL2 broke it, but WSL2 is exactly what I needed to match my use case.

> XAMPP did not work out of the box with me on Windows (skill issue on my part, I know), so my preferred setup was to run a Ubuntu Server VM (LAMP stack) and then develop whatever I had on a Windows IDE.

Why wouldn't you have just spent 5 minutes to get XAMPP working?

It's really a skill issue on my part.

LAMP stack worked for me perfectly on Linux out of the box, whether Ubuntu Server or any RHEL-based distro (even with SELinux enabled!).

I spent some solid 8+ hours on that, saw it uneconomical and went the VM way.

I stopped using VMware Workstation when WSL2 broke it

Is it still broken?

Nope, VMWare added the capability to work as a sort of nested hypervisor atop Hyper-V (which WSL2 and newer Windows security features depend on).

That being said, there is a performance impact.

WSL gave me the push to switch from macOS to Windows. And I couldn't be happier, tbh. There was a lot lacking in my Hackintosh/Windows dual boot setup.

> Edit: for clarity, by "multiple OS" I mean multiple Linux versions. Like if one project has a dependency on Ubuntu22 and another is easier with Ubuntu24. You don't have to stress "do I update my OS?"

For this part, I just create systemd-nspawn containers.

Last time I wanted to test something in a very old version of WebKit, creating a Debian Jessie container takes a few minutes. Things run at native speed.

You use distrobox (https://distrobox.it/) and move on with your life. At work I use multiple versions of Ubuntu seamlessly without messing with VMs on a host fedora box without issue. That includes building things like .deb packages.

> Like if one project has a dependency on Ubuntu22 and another is easier with Ubuntu24. You don't have to stress "do I update my OS?"

Have you tried lxd? It's far less janky than Docker (IMHO) to achieve what you describe. Docker is uniquely unsuited to your use case.

I love WSL, but you can do these things with Distrobox.

I'm with you - after years of messing with dualboot Linux, including (foolishly) running multiday Gentoo builds, WSL + Windows now gives me everything I want from Linux with zero friction.

In fact, I'm a little annoyed that I can't get a comparably smooth experience on my MacBook without spinning up a full QEMU VM. I know it's a bit hypocritical since, like most people, I run WSL2 (which is container/VM-based), not WSL1 (the original magic syscall translation vision).

Does anyone know why there's no lightweight solution on macOS - something like LXC plus a filesystem gadget - that would let me run stuff like "apt-get install chromium"?

Try https://tart.run/

>Native performance Tart is using Apple’s native Virtualization.Framework that was developed along with architecting the first M1 chip. This seamless integration between hardware and software ensures smooth performance without any drawbacks.

> WSL1 (the original magic syscall translation vision).

Actually, the OG "magic syscall translation" is Cygwin[0], which dates back to 1995[1].

[0] https://cygwin.com

[1] https://en.wikipedia.org/wiki/Cygwin

Edit: Fixed prose.

Absolutely! I remember playing and struggling with Cygwin back in the day… I meant original in the sense of the original vision for WSL.

Perhaps with the Mac Hypervisor someone is working on it.

But Qemu (via UTM) starts up pretty quickly for me. No slower than WSL2 under Windows. My only issue is that it seems to drain power even when idle.

Is this close enough? https://github.com/lima-vm/lima

I think WSL is great but if your only goal is to run several Linux OSes, any hypervisor will do. I think Proxmox is better suited to your use-case (hosted on Linux).

I love WSL because it lets me have the best of Windows and Linux.

Is it not the case that wsl2 is a vm; it requires hyperV enablement; and that turns your main windows OS into effectively a type of privileged vm, since hyperV is a type 1 bare metal hypervisor?

This is not often discussed, so it took me a lot of digging a couple of years ago, but I'm still surprised this is never discussed as a consequence / side effect / downside of wsl2. There are performance impacts to turning on hyper V, which may or may not be relevant to user (e.g. If this is also their gaming machine etc:)

You don't stress about Windows updates? Hard to believe it.

Yeah exactly ... I want Windows running in Linux, not the other way around, so I actually control the software and the updates!

I actually just tried WINE for the FIRST time (surprisingly, I have been out of the Windows world for so long)

https://www.winehq.org/

And as long as I installed the binaries from their repo, not Debian 12, it worked very well

Wine is an impressive project too. It's not a VM, which has upsides and downsides, but I was able to run GCC-TDM, Python 3, and git bash in it!

What do you mean by that?

As a reply to: You don't have to stress "do I update my OS?"

> It's an absolute delight to use, out of the box, on a desktop or laptop, with no configuration required.

I have been using it since the beginning of WSL 1 with a very terminal heavy set up but it has some issues.

For example WSLg's clipboard sharing is buggy compared to VcXsrv. It doesn't handle pasting into Linux apps without introducing Windows CRs. I opened an issue for this https://github.com/microsoft/wslg/issues/1326 but it hasn't gotten a reply.

Also, systemd is still pretty sketchy. It takes over 2 minutes for systemd services to start and if you close a WSL 2 terminal for just a few minutes systemd will delay a new terminal from opening for quite some time. This basically means disabling systemd to use WSL 2 in your day to day.

Then there's this 6 year old issue with 1,000+ upvotes https://github.com/microsoft/WSL/issues/4699 around WSL not reclaiming disk space. It means you need to routinely shut everything down and compress your VM's disk or you'll run out of space.

Beyond that is does work well so I'm happy it exists.

That doesn't sound good. I was planning to set up a Windows/WSL2 box, but this gives me second thoughts. Where can I read more about this?

It's still ok even without systemd. Technically systemd is disabled by default, you have to turn it on with systemd=true in /etc/wsl.conf.

I can't find a definitive source with an open ticket but if you Google around for "WSL 2 systemd delay startup" you'll find assorted folks talking it about with a number of different reasons.

I just went by my end results of there is a delay with systemd enabled and no delay with it disabled.

never had problems of systemd/2 minutes delays

not sure what would be the correct test here, but:

root@LP-T16:~# uname -rn

LP-T16 5.15.167.4-microsoft-standard-WSL2

root@LP-T16:~# time systemctl restart ssh

real 0m0.039s

user 0m0.008s

sys 0m0.001s

The delay is related to starting WSL 2, not starting a systemd service btw.

Maybe it's specific to Windows 10 Pro, who knows. I'm using the latest WSL 2 from the MS app store.

I just know when I installed Docker directly into WSL 2, when I launched a terminal I could not run `docker info` and connect to the Docker daemon for 2 minutes. The culprit was the Docker service was not available. I was able to reproduce this on Arch and Ubuntu distros.

Separate to that systemd also delayed a terminal from opening for ~15 seconds (unrelated to Docker).

After ~10 minutes of the terminal being closed, both issues happened. They went away as soon as I disabled systemd.

Jumping on the anti-wsl bandwagon; I just can't abide the loss on control on windows, will the next update ignore/reset/override my privacy settings? What Gordian knot must I slay to have a local only account (Thanks Rufus!) How do I turn off/uninstall a million things I don't want, Xbox game bar?!?

Linux or *BSD give so much more respect to the user, on windows you are the product! Stand up for yourself and your data!

It doesn't work on any of my 3 Windows machines, all completely different hardware. Jank factor 100% for me. I wish I was seeing what you're seeing.

I like that wsl is a thing when I'm on a windows machine, but it can also serve as a reminder of the often unnecessary frictions that exist between operating systems.

When the answer to a "how do I do X on windows" question begins with "start WSL", my primary reaction is frustration because they're basically saying "there's not a good way to do that on Windows, so fire up a Linux VM".

Just to pick my most recent example, from today. I wanted to verify the signatures on some downloaded rpm files, and the rpm tools work on linux. I know, rpm files are native to a family of linux distros, so it's not surprising that the tools for retrieving and verifying their signatures don't work on windows but... it also seems reasonable to want a world where those tools can install and run on windows, straight from a PowerShell session, with no VM.

Multiply that by all the little utilities that that can't be deployed across multiple operating sytems, and it just seems like some incompatibility headaches are never really going to go away.

Still somewhat janky. I use it on my work machine (since it at least seems a bit faster than using VirtualBox) and regularly run into issues where npm won't build my project due to the existence of symlinks [1,2]. wslg windows also don't yet have first-party support from the windowing system [3]. I also remember having trouble setting up self-signed certs and getting SSL working.

1. https://stackoverflow.com/questions/57580420/wsl-using-a-wsl... 2. https://github.com/microsoft/WSL/issues/5118 3. https://github.com/microsoft/wslg/issues/22

Now if they could only do Windows 12 by taking baby steps in yearly release of Windows 11.1, 11.2 etc.

Iterating on improvements and polishing on Screens and Design that they haven't touched in the past 30 years. Improving on ARM support etc. And STOP adding Ads on the OS.

And the Surface Laptop continues to push Hardware quality forward. From Speaker, Touchpad, Screen, Motherboard etc.

It's not literally true that "Weasels Ripped My Flesh"[1] but WSL2 did rip the python support in QGIS by polluting my PATH with space characters.

[1] https://en.m.wikipedia.org/wiki/Weasels_Ripped_My_Flesh

I like WSL for this single reason too - it gives me space to run isolated experiments without touching my primary OS. So if that's what windows users get out of it, cool.

You can do the same thing with many other technologies on most other operating systems. I've used, in chronological order: FreeBSD jails, VMs, Cloud-hosted VMs, Docker, K8s, and Nix flakes. WSL is probably somewhere in around K8s.

My point is, we've had the ability to run "subsystems" for decades, by different names, on every OS. WSL is cool but quite late to the game, far from being "more powerful than linux".

Perhaps "more powerful" is also a factor of who is the computer user. For example, Linux is not as "powerful" if the computer user is someone who knows little about how to use it.

For a person who will not invest the time to learn, e.g., how to avoid or minimise dependencies, indeed something like Windows with WSL may appear "more powerful".

The point of this comment is that "power" comes from learning and know-how as much as if not more than simply from choice of operating system. That said, some choices may ultimately spell the difference between limitations or possibilities.

Install Promox or TrueNAS on a bare metal desktop to experience the true power of multiple operating systems running simultaneously. On most days, I am running multiple VMs with these OSes in parallel: Windows Server 2025, Windows 11 Pro, and these flavours of Linux - TrueNAS/Debian, Ubuntu, Manjaro, Zorin OS. I also have a dozen or more lightweight containers running, some with LXC on the bare metal host and others with Docker inside the TrueNAS VM.

This setup automatically backs up my data and is resilient to disk failures. It’s the ultimate form of power and bliss.

I used to agree with this for WSL1. Syscall translation gave solid performance, decent FS integration, and interop within WSL with windows executables. I really liked it.

WSL2 has been such a pain. You're basically managing a VM with VMWare Tools somewhat more integrated. I gave up on WSL2 after a few months and went back to booting my arch installation most of the time. Now I'm on a mac for the first time in a long time because windows has gotten so bad.

This is doubly sad because the NT kernel is so well designed to host multiple OSes due to the OS/2 stuff decades ago. All wasted.

It is really good but honestly would prefer something a little more like:

- Linux that works great on a laptop / does the right thing when closing the lid - Linux that doesn't have worse battery life than Windows / macOS - Seamlessly runs Windows when you need to run something (e.g. click on Excel) - Isn't necessarily free (prefer quality over low price in this situation)

Windows of course has many of these traits and WSL is a pretty good compromise, but I would prefer to boot into Linux and use Windows only when necessary (since my need for it is less common).

WSL is great if you're on Windows, but I wouldn't say it's more powerful than Linux. Distrobox on Linux covers your "multiple OS" use case quite well.

Windows treats you like a baby. You cannot learn the internals of it and it forces decisions on you. With Windows, the computer that you paid for is not yours.

You can run multiple OSes simultaneously on Linux itself - Linux can run VMs just fine. I.e. Linux guests on Linux host and so on. Take a look for example at virt-manager (libvirt / qemu + kvm).

And WSL is a limited VM using HyperV anyway. If you want to run a VM, you can as a well run a proper one which isn't limited and runs a full blown distro with sane configuration.

So WSL is definitely not more powerful than normal Linux.

And yet when I reboot my computer windows has shown me an entirely new place I can see ads - this week it was my lock screen.

So I left - I am willing to do more work to be spied on less, to be used as a product less, and to fight with my computer about who owns it less.

> and to fight with my computer about who owns it less.

This is a great way of saying it and expresses the uneasy feeling windows has given me recently. I use Linux machines but I have 1 windows machine in my home as a media PC; and for the last several years windows has made me feel like I don’t own that computer but I’m just lucky to be along for the ride. Ramming ads on the task bar and start menu, forcing updates on me, forcing me to make a Microsoft account before I can login (or just having a dark UI pattern so I can’t figure out how to avoid it, for the pedantic).

With Linux I feel like the machine is a turing complete wonderbox of assistance and possibility, with windows it feels like Microsoft have forced their way into my home and are obnoxiously telling me they know best, while condescendingly telling me I’m lucky to be here at all. It’s a very different feeling.

Yeah, "Weather and More" is such a joke. I like the idea of Weather on my lock screen in theory, and I sometimes miss Windows 8's great support for Lock Screen live data, but I have huge problems with almost everything else in the "and More" (news, no thanks, ads, definitely no thanks, tips, maybe not). Thankfully it is still really easy to turn off "Weather and More", but I wish they'd give us a "Weather and Nothing Else". (Same reason one of the first things I do is disable the "Widgets" display on the taskbar in Windows 11. Weather is great, everything else I don't want and/or actively hate.)

Yeah this is what pisses me off the most about windows. Telemetry that can't be turned off normally. Ads everywhere. Microsoft deciding when I must restart for updates. Microsoft trying to manage my behaviour telling me to try new features. Screw that. My computer is my own and must do what I choose.

This feature thing is really one of their strategies. At work they send us "adoption managers" that run reports to check whether people use feature xyz enough and set up stupid comms campaigns to push them to do so.

I really hate that. I decide how I use my computer. Not a vendor.

The development experience is relatively cumbersome compared to using a native Linux distribution and containerizing application dependencies where needed.

Last time I used it they kept hogging some common keyboard shortcuts for whatever Windows stuff even though the VM-window was focused. Did they stop that?

I've lived in WSL for 3 years now, and have zero complaints. It has worked with no issues what so ever. In 2025, Windows is the best Linux UI.

Using WSL on Win11. I would prefer Linux but I never got used to Open Office/Gimp/... and need to use PowerPoint / Affinity. But WSL mostly works, and added some tools and config to make it useful with WezTerm

https://www.amazingcto.com/upgrading-wsl-with-zsh-and-comman...

You can run multiple Linux distributions in chroots or containers, such as docker containers. I have showed people how to build packages for Ubuntu 22.04 on Ubuntu 20.04 for example.

This is what tools like toolbx or distrobox solve. You can have easy to use containers with libs from any distro with a few commands, using podman or docker as the backend.

It's a... VM? Like the Linux VMs running on Linux computers in the cloud?

Sorry but not sorry, it's not easier to run than on linux. It requires the Windows store to work, and to use Hyper-V (which breaks VMware workstation, among other things).

It's in a better package, to be sure, but it's not "easier to run multiple OS on the same computer". It's easier to use multiple OSes (no SSH, GUI forwarding, etc), as long as all those OSes are Linux flavors supported by WSL.

Want FreeBSD or Windows? Nope!

Does it really need the store? I thought you could just go "wsl install" on the console.

The files, including and especially the distro files, `wsl install` installs still originate from the Store's CDN, so the truly paranoid that distrust the Store (including some corporate environments) and just entirely block Store CDN access at the DNS and/or firewall level still break WSL installs.

You're likely right, I haven't used it in ages. Though I recall that at one point you had to get distributions from the Store, but it may have been that long ago that it was still being called "Bash for Windows".

You don't need the store.

> Want FreeBSD or Windows? Nope!

Well, it is windows subsystem for Linux :) not windows subsystem for windows or FreeBSD for that matter :)

Ps I wonder if you can make your own image? After all its really just Hyper-V with some config candy.

It's a bit more than just some candy, there's substantial glue on both the Linux/Windows sides to get Plan9, WSLG, and the other components to work.

That said, the kernel they distribute is open source and you're not limited to just the distros they're working with directly. There are a number of third party (e.g. there's no Arch from Arch or Microsoft, but there's a completely compatible third party package that gives you Arch in WSL2)

Haha yes, I was being cheeky :)

I'm pretty sure that with the opensourcing, we'll see freebsd or more exotic systems popping up quite quickly. Heck, macOS would be fun!

You're right, it is incredibly nice. Just the other day I got a Windows-only developer to install and use the POSIX/*NIX toolkit we use for development/deployment. In 30 minutes he was editing and deploying left and right with our normal open source stack. No messing around with Cygwin or MSYS or anything, it all just worked in Ubuntu on WSL. It's fantastic.

I share your sentiments. Makes testing my builds against windows, Ubuntu 22, Ubuntu 24, etc a breeze. It pretty much 'just works' and I can take it to go on my laptop. Even though I do most my work in Linux, Windows is a convenient 'compatibility layer'. I was skeptical at first when my friend suggested I try this, but daily usage has won me over.

WSL is massively slower than Linux. Not just the 10% or so for VM, but probably 50-90% slower for disk access. It takes many times longer to start tmux. It has update bugs that crash open terminals and that's not even part of the regular windows forced-update fiasco. In short, it's garbage. It's one of the primary reasons I moved back to Linux for my daily driver.

I'm old enough to remember that before docker there was chroot. It's fairly easy to put lots of different user mode portions of Linux distros into directories and chroot into them from the same kernel. It seems a bit like what you're asking for.

There's also debootstrap which is useful for this technique, not sure if it also works on Ubuntu.

debootstrap absolutely works in Ubuntu

For WSL 1, I kinda agree. It was basically the Posix Subsystem re-implemented and improved. Technically amazing, and running parallel to Windows without virtualization. Too bad it had so many performance issues.

But WSL2 is just a VM, no more, no less. You can do the same with VMware Workstation or similar tools, where you even get a nice accelerated virtual GPU.

My only big gripe with WSL right now is GUI applications. wslg is not good, and the only good experience is when applications have a good remote development UX such as vscode.

Another, smaller, gripe is networking. Because of how WSL is networked, I've run into edge-case issues with connecting to networked applications running in WSL from Windows.

You need to make sure that they use Wayland. Running X11 apps is significantly slower in wslg. Native Wayland apps run much faster.

Run a rootless X server (XWin, Xming) on Windows, network the two (SSH tunnel), you have GUI Linux apps on Windows.

Lack of all packet types disqualified it for me. Is there any hope for nmap, etc?

I use WSL, but I'm actively looking for a way to move away from it. The only thing holding me back are languages like Ruby or Python, which are designed to work in a Unix-like environment. I briefly considered forking Ruby and stripping out all of the Unix-isms but in the end I gave up and just installed Linux (WSL).

docker is pretty easy to use on linux (even rootless docker isn't particularly painful) and KVM using QEMU is also pretty easy for running Windows things. I used WSL quite a bit but ultimately have switched back to running Ubuntu as my main.

Here's the main difference between making Windows vs Linux the main OS from my POV: Windows is a lot of work and only the corporate editions can be converted into not-a-hot-mess-of-distractions (supposedly). Out of the box Linux doesn't have all of the bullshit that you have to spend time ripping out of Windows. You can easily re-install Linux to get the "powerwash" effect. But if you powerwash Windows you have to go back and undo all the default bullshit again.

Having said that Windows+WSL is a very nice lifeline if you're stuck in Windows-land. It's a much better combo than MacOS.

WSL gives you no support for USB devices, which is a massive pain for embedded development when IT forces you to use Windows. Also, this might just be specific to my setup but WSL networking is very finicky with my company's VPN, and breaks completely if the VPN ever drops out requiring a full reboot.

WSL2 can forward USB devices

https://learn.microsoft.com/en-us/windows/wsl/connect-usb

I regularly run ADB through WSL2 using this.

For me it was slow, full of compatibility issues, and glitchy. Some simple packages wouldn't even install in the official Ubuntu WSL distro. To be honest I don't know what the use case for this is, other than to run some one-off Linux thing once in a while without having to use another box.

How long ago did you try that?

I use WSL2 to handle Linux (and Windows cross-) compilation regularly, along with running a number of native tools that are specific to Linux.

I've never had any issues with that, even to the point that I've been able to run MAME natively from Linux and have it show up like any other windowed app.

I agree with your opinion on WSL. I psy a similar "tax" when I defend ChromeOS, and I will not stop it, like you won't.

The Linux on Desktop is finally approaching, in more than one "shape", none of which is the shape some people expected/wanted.

Windows 10 with WSL(2) is/was peak Windows for me. You could build stuff and edit MS Office documents in the same place. Sadly, it wasn't meant to last. I have no intention of giving W11 a try, not yet decided what I'll be using come this fall.

I'll second you, WSL makes Windows a first class experience because now I can seamlessly have Linux and Windows apps in one laptop. Yes, I could run VMWare Workstation or HyperV, etc, but this is just better integrated.

I'm a daily driver. It completely changed the way I work. Am I curious if something will compile? Open a terminal and type make. The files are all already there. You can even run graphics apps. It's wonderful.

As of a couple of years ago the integration was not that great and I switched to just using a full-fledged VM instead. For example, trying to use binaries in WSL from within Visual Studio or vice versa was not great.

> WSL is more powerful than Linux because of how easy it is to run multiple OS on the same computer simultaneously.

I do that with KVM too, and each has their own kernel, not one shared kernel made and controlled by one vendor.

I heart WSL. Years ago I was going to switch to MAC OS to have a more unix like experience/workflow. Then WSL came out and I stayed because Linux is the environment I spend most of my time in.

> Like if one project has a dependency on Ubuntu22 and another is easier with Ubuntu24.

Sounds like you could benefit from Qubes OS, which runs everything in VMs with a great UX. Including Windows.

I agree it is a convenient way to run multiple Linux VMs, but it comes with the drawback of having to use Windows, which is a major impediment to anything I may want to do with my computer.

I use Ubuntu 22 in a LXC container on Ubuntu 24 (because of Webex).

I also run other Linux instances with KVM.

I even run a Linux x86_64 executable on an ARM SBC using QEMU.

I just feel that Linux is so much more flexible than Windows.

The power of linux with the professionalism of paid MSFT engineers

You can run multiple linux distros on linux just fine via KVM/QEMU, there is nothing special WSL offers except that it is a must if you're doomed to use windows.

qemu on Linux solves a bunch of these problems as well. But yeah, UX-wise WSL is pretty good at solving the problem of “provide Windows devs a POSIX environment”.

Qemu is nothing like wsl UX wise. The UX on windows is double click gimp and then a window for gimp opens. For qemu it opens a new window for the wm, has awkward input focus interactions, you probably have to log in to the vm, and it can not be easily setup to automatically open the app you want.

I used to love WSL when I had a Windows machine because I used lots of docker containers, but now that I am in a Mac with Apple Silicon, there is no going back.

ELI5 does it allow me to run windows programs in Linux?

WSL sucks and I much prefer having a true VM in Hyper-v. WSL is full of weird behaviour and gotchas. Docker?? no pid 1, weird kernel etc ...

>WSL is more powerful than Linux because of how easy it is to run multiple OS on the same computer simultaneously

Is VMWare more powerful than Linux?

Previously, I had dual boot with ubuntu and windows. Sometime last year I just removed ubuntu, and haven't regretted it.

wsl works good enough.

I still have issues with the networking but I agree. Its a fantastic system and it shits me only that it could be a bit better.

It’s a delight to use if you don’t mind your computer conducting 24/7 surveillance on you for a multinational corporation.

If you want to “run multiple versions of Linux at once” and don’t like plain Docker, maybe check-out Podman Desktop.

Most people have little use for running multiple OSes, and that drops a lot when you just abandon Windows entirely.

I want to know what limitations and tradeoffs am I embracing when using WSL vs booting linux off a usb stick.

On linux, I've been using lxc (now incus) for years to get different distros.

Can do the same with FreeBSDs Linuxulator. I run Arch Linux on FreeBSD, emulated.

WSL is so incredible. But support for it from 3rd party dev tools is so terrible.

I agree with you. Maybe if it had AI shoehorned in, hn would be happy.

This would be a great point if WSL didn't require running Windows

I tried it and found it to be such an abomination. I can’t understand why any self respecting software developer would use Windows with a bastard linux like WSL instead of just using actual Linux. Feels like a massive skill issue.

it's for when your corpo provides you with a windows laptop for development. hope that helps with your understanding

Well I guess now you just need to add WSL support to wine.

You can do the same on Linux. Distrobox exists.

We already have Distrobox that does same thing.

> Every time I praise WSL on hn I pay the karma tax

Hmm...

> WSL is more powerful than Linux

Oh.

I'm not the biggest fan of WSL2, but it's definitely good enough for people to like it. it's worked well enough for me in the past, but the last time I used it, there were problems with mDNS and BPF that it just made more sense for me to boot into leenucks.

But you're definitely not crazy for liking it. And people should chill out instead of downvoting for someone who just says what works for them.

I haven't tried Win11 and probably won't unless my employer forces me to. But if Win11+WSL2 works for you, more power to you.

Have you used Distrobox?

I won't downvote you, but I will die on the other hill - the one over there that has a guy sitting down with his arms folded sporting an angry face every time someone something positive about WSL. There's at least three of us on that hill. And we're not going anywhere.

I'll second this, and I'm someone who ran a certain alternative OS to Linux before Linux was viable instead of run Windows, worked as a developer of Win16 and Win32 apps early in my career which gave me a deep love-hate of the platform, couldn't stand Microsoft's monopoly tactics back in the 1990s and 2000s, and remain ever-sceptical of Microsoft's open source and Linux initiatives...

... but WSL is an excellent piece of work. It's really easy to deploy apps on. Frankly, it can be easier to a deployment there than on a Linux or macOS system, for example the reasons detailed above.

Real talk. And anybody who argues is taking a heavy dose of copium to justify their use of Linux and the ensuite of compatibility issues that entails. Let them have their sense of superiority :' )

... You know that you can run VMs, or full-OS containers on a Linux desktop right?

Or on a macOS Desktop. Bonus: doing so on either platform doesn't also mean your host OS is running under a hypervisor, as it does with WSL2.

Bigger bonus: you don't have to run fucking Windows.

> Bonus: doing so on either platform doesn't also mean your host OS is running under a hypervisor

Why do you think, technologically, this is some form of "bonus"?

Windows by default runs on a hypervisor since some Windows 11 version.

That may all very well be, but uuh, you're then forced to use Windows

> WSL is more powerful than Linux because of how easy it is to run multiple OS on the same computer simultaneously.

This is why you pay karma tax. This statement is so clearly representative of a falsity.

My linux can run multiple linuxes as well without VM overhead. Something Windows can’t do. Furthermore WINE allows me to forgo running any vm to run windows applications.

I developed on WSL for 3 years and consistently the biggest issue was the lack of ability to use tooling across the shared OSes.

Your karma depleting statements are biased, unfounded, and it shows as you do not really provide counter evidence. That’s why you lose karma.

Except Wine cant cover all of Windows (partly due to fault of Windows). I can't run UWP apps for example. Windows is not a good operating system but if you need it. WSL creates way more intuitive working environment for you. So even if you can run multiple Linux OSes in Linux you can't run Windows as easily you can do linux on Windows. So OPs statement is not incorrect.

I totally agree and will join you on the hill. I used Linux exclusively at my job for two years straight and now do the same job but from Windows 11 with WSL 2 on the same physical ThinkPad T41 laptop. Windows gets the basics right more than Linux did (sleep states, display, printing). And as the OP notes; it makes it easy to run multiple distributions and never fear that something I install or reconfigure within the WSL2 terminal will screw up my host. Having a different OS improves isolation in this regard, not at a technical level but for me making mistakes and entering commands in the wrong place, since Windows does not accept Linux commands. JetBrains and VSCode both have great support for WSL2.

Reply on news.ycombinator.com