Windows on Unix?

I was just wondering the practicality, pros and cons for Microsoft to go the route of Apple and use Unix to build their operating systems on.

Especially since Unix is praised to be so stable and secure and has gotten such positive acclaim for being the foundation of OSX, what can we say why or why won't Microsoft adapt it in their own OSes?

Thoughts?

 

You might want to look at Microsoft "UNIX"(fake) SVR.
No replacement for the real thing but it is designed to run UNIX apps.

 

As great as UNIX is, I think Microsoft hit a homerun when they wrote the NT kernel. It's a LOT more capable and "more future-proof" than people give it credit for.

While I don't know the specifics (I'm not a big enough nerd), lets face the facts. The NT kernel that gave us Windows XP, gave us Vista, gave us Server 2003, gave us Server 2008, gave us Win7, etc etc etc. The same kernel (well, roughly) that ran on friggin 386s now give us DirectX 11.

Sure. While it would be nice to rewrite it and go back to fix stuff, I think about the transition. Few people remember when OS X first came out. It was a nightmare. Developers cried. Users cried. It was like a funeral, except in aluminum (well, not really, but you get the picture).

 

Reason not to use UNIX
1)UNIX is a standard has to adhere to a specific standard not exactly Microsoft style.
2)Means people can switch to other UNIX more easily will spoil the monopoly Microsoft painstakingly built up.

I don't think you should confuse UNIX with Kernels.
From what I know UNIX means POSIX compliant they can use any kernel.
It can be Darwin Kernel (OS X), Solaris Kernel, Linux or BSD kernel (if we loosely generalize UNIX as POSIX standard not because developers pay monthly subscription to keep the label)

 

That's a valid point. I often confuse that myself.

And UNIX isn't the foundation of OS X. It's FreeBSD.

 

There is no reason for MS to switch to a completely dfifferent OS. So, no pros, but lots of obvious cons. End of story.

Unix systems are not inherently any more stable or secure than NT-based systems. As a matter of fact, NT has a significantly more sophisticated security structure. How that structure is being used, and why, is a different question altogether.

 

Microsoft does appear to be researching new kernels to use for future releases, including one project that involves running multiple kernels in parallel. You should just Google this stuff to find out more as it isn't a secret, but needless to say MS is not interested in using Unix as a base for future designs.

Unix is stable and reliable, but I'm sure there are other criteria to keep in mind when considering hardware and computing trends.

 

Not really
Foundation of OS X is MACH + 4.4BSD + Apple Propriety AQUA + FreeBSD userland.
4.4BSD is old code it is so old that FreeBSD already stopped using it.
They moved on and Apple still think they have the most advance OS
Microsoft used to use FreeBSD TCP/IP stack like many other OS they were widely used as TCP/IP was originally written for BSD.
However if I remembered they rewritten it in Vista.

 

yup, it's called KDE with Wine.

MSFT has already thrown out three commercial unix-like operating systems.

 

Well, Mach was designed for BSD. So ya. FreeBSD. Or purist will correct me and say it is just BSD. But whatever.

 

Hmm, let's consider the OP's initial questions:

Practicality: Near impossible, they might as well write a whole new OS
Pros: Stability... Probably other things that aren't immediately apparent
Cons: It would break compatibility every single Windows application in existence - I'm sure Apple and the Linux people wouldn't be able to contain their enthusiasm if MS did this though, albeit for different reasons

Let's just be happy that we're not using DOS-based Windows anymore. The NT kernel may not be perfect, but in practice, it's good enough for modern computing.

 

Compatibility and the effort required make such an effort all but pointless. Add to the fact that NT is working perfectly well for Microsoft, and why would they try it? Mac OS was failing from a business perspective when Apple changed things up drastically with OS X - I'm sure if Mac OS 9 had been thriving they wouldn't have gone to such lengths.

More likely is that Microsoft would add compatibility for Unix programs to Windows. The NT kernel actually had some built-in Unix compatibility at one point, as well as limited OS/2 compatibility (which was removed in XP). With the success of Windows, OS/2 and Unix compatibility on NT proved unnecessary, and development on OS/2 and Unix compatibility on NT all but stopped.

 

That's the beauty of the NT kernel. Adding support is only a matter of adding modules, rather than complicated rewrites of entire sections of the kernel.

Edit: This brings me back to those Vista threads calling Microsoft to write a new kernel. While I'm aware that the OP isn't advocating such an action, I often question such large scale, radical steps that doesn't solve the core issues. Rewriting a kernel is an immense undertaking requires millions and millions of manhours. Not a smart business tactic when probably 90% of the kernel works perfectly. WHy throw out the baby with the bathwater?

 

Don't worry it is the irrational anti-Windows Apple fans making their rabid remarks.
They like to reinvent the Wheel.
Speaking of which there is MinWin a proof of concept Windows modular OS.

 

It's not a proof of concept. It's more like an organization chart. It's inside Windows 7, and in Vista. And in Server 2008.

They basically made a huge chart of all the components in Windows and "arranged" them to where components close to the core of the kernel are only dependent on components below them. It's not a proof of concept like Midori.

This is one of the biggest reasons why Server Core exists. With a well defined set of guidelines and organization, they can draw a line across the chart and say "This is the absolute minimum required components needed to meet this specification."