Windows 10 Gets A New Linux: openSUSE (fossbytes.com)

An anonymous reader writes: "Running Linux binaries natively on Windows... that sounds awesome indeed," writes Hannes Kuhnemund, the senior product manager for SUSE Linux Enterprise. He's written a blog post describing how to run openSUSE Leap 42.2 and SUSE Linux Enterprise Server 12 SP2 on Windows 10, according to Fossbytes, which reports that currently users have two options -- openSUSE Leap 42.2 and SUSE Linux Enterprise Server 12 SP2. Currently it's Ubuntu that's enabled by default in the Windows Subsystem for Linux, although there's already a project on GitHub that also lets you install Arch Linux. "It's quite unfortunate that Microsoft enabled the wrong Linux (that's my personal opinion) by default within the Windows Subsystem for Linux (WSL)," writes Kuhnemund, "and it is time to change it to the real stuff.

Comments

Re:Running Linux on Windows is awesome? How so?

[dunkelfalke]

Doing what I do now - developing for Linux in Visual Studio. And, to be honest, even though I develop for Linux, I personally prefer using Windows on the desktop both at work and at home (my little home server runs on Debian, but it is mostly used as a data graveyard and the only time I actually use it is when running midnight commander in a ssh session).

Why is that useful?

[cjonslashdot]

Why run Windows in the first place? I am an Agile transformation coach, and I work in large organizations, and I always wonder, Why, if they are deploying on RHEL, are their developers writing code on Windows laptops? The problems that result are endless. And the solution is simple: either (1) run real Linux in an VM; or (2) run Linux natively. #1 will satisfy enterprise access to email, etc. The solutions are already here. Trying to cram Linux into the Windows kernel seems bizarre to me. What do others think?

Re:Why is that useful?

[ledow]

Embrace.

Extend.

Extinguish,

They're hoping that "linux" comes to mean just a particular set of utilities, no matter the OS.

In this day and age, virtualise. And it doesn't matter what OS you host virtual machines on, so long as they run.

Which is a death-knell to Windows, because the choice between "server core" and a barebones Linux install with a hypervisor? What's to choose except price and licensing?

Developers should be able to code on - literally - anything they want to. It helps in testing, if nothing else, if they are checking in code that is Windows-only and everyone complains that it breaks builds.

But they should all have all the target platforms as VMs, too. Then it's a matter of personal preference.

To be honest, I don't get why so many coders actually use MacBooks. It seems completely the wrong decision to me, if given free choice.

But the days of which OS is actually running on the hardware mattering are long gone. The choice of what you use as desktop is personal preference. The choice of what to use for backend services doesn't matter so long as you have people managing it.

Windows, at this point, is just a fancy GUI, not unlike which choice of DE you use on Linux. I think Microsoft are trying to claw that back a little and make you think that you can get rid of the Linux desktops and interfaces by using Windows.

Re:Why is that useful?

[swb]

You said it yourself -- "large organizations".

They're aiming for some kind of economies of scale in purchasing, application deployment and security that go way beyond the single-digit percentage of user base that developers represent.

They could just hand over the hardware and let the developers run their own machines, but this has all kinds of security implications and often bleed developer productivity in desktop maintenance overhead.

Running dev machines natively in Linux makes some sense, but may cut them off from other Windows-only applications they need to be part of the larger organization. as well as lack of visibility in enterprise management software. Running it in a VM has the same problems plus the added complexity of two environments.

I doubt Microsoft's solution is designed principally as a developer solution, either, but probably a long-term gambit to make it a more universal platform to retain users when the year of Linux on the Desktop rolls around. They must see some future in their crystal ball where enough Linux desktops exist that *not* being able to run some application is an existential risk to Windows.

Re:Real Stuff

[Anonymous Coward]

Weird, isn't it. Getting Linux working on Windows has to be MS's effort, since they're the ones with access to the build process of official releases, and they have access, like everyone, to Linux code and build process, even for SuSE. Getting Windows on Linux has to be MS's effort, since they're STILL the only ones with legitimate access to the source code of Windows and the "patented" stuff therein.

Yet it's Linux's fault that Windows programs don't work on Linux, and Linux's fault that Linux programs don't run on Windows.

MS still have the only legitimate access to all the information necessary to make this work. But failure is someone else's fault...

Re:Shudder.

[JanneM]

Sounds horrible to me. Why bother?

Not sure what MS' motivation is, but it's good news for a lot of scientific software developers. Small teams or single researchers rarely have enough time to even keep the main development going, never mind keeping up with multiple OS targets. With this everybody can simply focus on Linux, and tell Windows users to just run it under the Linux layer and stop asking about a native port.

Re: Running Linux on Windows is awesome? How so?

[Joce640k]

Well, uh...

"sensible" is the word you're looking for.

Visual Studio beats anything on Linux. If that's your main use for a computer then run whatever system it takes.

Said before, but bears repeating

[jenningsthecat]

Linux on Windows is part of Microsoft's 3-E strategy. If they can stunt the growth of Linux as an OS by co-opting Linux applications to run on Windows, they may eventually succeed in cutting the heart out of FOSS altogether. And they would LOVE to do that, because FOSS is one of the few significant forces standing between them and the conversion of the whole world to a software-as-a-service model, wherein the average user doesn't own shit and has fuck-all in the way of rights, choice, or legal recourse.

Anybody who has a choice shouldn't run Windows, and certainly shouldn't run Linux applications on Windows. And anybody who MUST run Windows, should also run Linux, and use Windows ONLY for those things that absolutely require it.

Re: Real Stuff

[Rob Y.]

People run RedHat for the long-term support. Enterprises don't like being forced to upgrade on a vendor's schedule, and RedHat was the first Linux provider to recognize that and cater to it. Timely security upgrades for a consistent platform - over years - is what enterprise users want. And like it or not, that is a technological meaning.

Yup, GNU/NT-Kernel

[DrYak]

If I understand it right, it's a GNU/Linux distro without a Linux kernel on top of a compatibility layer on Windows, right?

Yup, mostly(*).

So "GNU/Windows NT Kernel" is better than "Linux" - That actually one of the rare few occastion a typical "GNU/Linux" distro gets used without the Linux kernel part.

But because "Linux" has brand recognition, it's still used.

---

(*): there's no separate compatibility layer (unlike things like Cygwin which are a user-mode compatibility layer that translates POSIX API-calls into Win32 calls - and thus enables soure compatibility).
The NT-Kernel has a bizare peculiarity : it can export several different ABI's to usermode software - it has different "personnalities".
- Win32 is just *one* of the set of ABI available.
- A long time ago, that made it possible to run OS/2 software on Windows NT.
- A little bit less longer time ago, Windows NT also had a "Unix" personality.
- Now WSL is actually the NT kernel exhibiting a small subset of the ABI featured by the linux kernel - about the bare minimum to get a few basic user-mode software (e,.g.: the "GNU" part of "GNU/Linux") run unmodified.

These are straight ABI available from the NT-Kernel, not a mere Linux-to-Win32 API conversion like Cygwin.

e.g.:
- Among other defaults Win32 has a poor multi-processing (forking is expensive). Cygwin application have to rely on that poorer cousin in order to provide multi-processing to POSIX.
- The recent kernels of Windows NT intoduced pico-thread which are very cheap, weren't available in the Win32 API back when introduced, but where exposed through the "Linux-lite" API that is WSL in order to make a usefull multiprocessing.

On the other hand WSL is far from complete. There is tons of stuff that you can do on your GNU/Linux that you can't do with WSL (e.g.: filesystem drivers)