* Smooth set up. Don't have to install some large commercial 800MB MSI like VMware workstation, download some Linux image, go through partitioning of file system etc.

* Well integrated. I can open up a terminal and it acts as any other window in my system (meaning I don't get the window in a window effect as you get with a new VM).

* My file system is mapped automatically. No need to set up Shared Folders or whatever manually.

* Better startup perf. WSL starts in a second on my computer. Never had the same experience with full VMs. Even if I use something like alpine just starting VMware or VirtualBox takes a lot longer than starting WSL.

Saying it is like any other VM seems just incorrect. Saying it didn't work out seems even more misguided.

That may be, but the rest of your comment seems to be unrelated to the matter at hand since you merely listed some advantages of WSL2 instead of addressing the disadvantages that are causing people trouble.

Many of them have ports to Windows but it's just easier when it's all already available. Some days ago I needed to run some penetration test software and while supposedly I should be able to download the code myself and build in Windows, just installing using apt is a lot easier.

https://bbs.archlinux.org/viewtopic.php?id=204875 2015 November

https://bbs.archlinux.org/viewtopic.php?id=206032 2015 December

https://bbs.archlinux.org/viewtopic.php?id=210685 2016 March.

And these were times when I needed community help, most of the time I could get it working by pairing again or some such nonsense. It never worked reliably, in general.

Note I switched to Windows as my daily driver in 2018 January.

It seems sound is still a gigantic mess of PulseAudio and ALSA https://wiki.archlinux.org/index.php/PulseAudio

So even Ubuntu isn't necessarily a guarantee of stability.

Bluetooth has never worked.

Most decent languages work great cross platform. I'm not sure what I'm supposed to be missing.

To be frank, the only think that bothers me about web dev on Windows is that the Elixir repl doesn't support "up" for editing the previous command.

Most people choose their laptops based on the OS, not on the pointing device.

However my T420 and X220 still perform excellently even after 9 years.

2006 was for me the turning point where Linux was only used via VMs or some remote box.

WSL2 is just fine, I don't need to mess with VMWare any longer.

And if I am not doing Linux deployments, I can do plain Windows development, like I have done since Windows 3.0.

For the Linux desktop experience I have a surviving ASUS netbook.

This similitude baffles me. Do you mean "had to wear protective equipment, had to drive constantly between locations, we were knocked over every few seconds, and we risked drowning several times"...?

Maybe there was also some long term goals, like making Windows better platform for Linux software in general (server use).

It's made mine and my coworkers lives so much easier compared to before, so I'm certain use cases exist. Since WSL2 runs on HyperV, for some of my coworkers it's not an option, since they rely on VMware and similar for other essential work.

Tell me which you would rather use for node and or docker development.

- com.docker.hyperkit is nuts. I can have a single container idling in the background doing nothing, hear my laptop fans spin up, and know without checking that com.docker.hyperkit is using 200% CPU for no discernible reason. Restarting the daemon brings things back down...for a while.

At the office we use Linux, but since I've been working from home on my Mac I've gone back to Vagrant for as many projects as I can. Heavier and far less easy to orchestrate yes, but I've found I actually end up with better and more predictable performance (and far less time with my laptop doing double duty as a space heater).

I've been running this set up for over a year and the volume performance is superb.

Flask, Phoenix, Rails and Webpack driven apps are all a fantastic experience. For example Webpack takes 150ms to compile SCSS / ES6 JS diffs for large real world projects. Web server reloads on code change are effectively instant and I get microsecond response times in some Phoenix apps in development.

This is on 6 year old hardware too and the source code isn't even sitting on an SSD (but Docker Desktop is installed on an SSD).

In all cases, everything is running in Docker through Docker Deskop and I use the Docker CLI / Docker Compose in WSL as a client to connect to Docker Desktop.

It's all documented at https://nickjanetakis.com/blog/setting-up-docker-for-windows....

I have WSL1 installed, but very seldom feel the need to actually use it.

These days, there's the Windows Terminal, which is fine, surely, https://www.microsoft.com/en-us/p/windows-terminal-preview/9...

Your statement goes into both directions.

Guess why "Year of Linux Desktop" has failed to happen, and no, ChromeOS and Android aren't really GNU/Linux, the kernel is irrelevant to userspace languages and public APIs.

Not sure why you'd use or mention 'MS-DOS' since that stopped being a thing about 20 years ago, a decade before node existed.

Big companies that employ lots of smart people frequently seize up and become incapable of innovating because their internal parts mesh against each other and halt the whole machine. I suspect that's what happened here.