I was surfing the 'Net and I came across this article from '06:
http://macdailynews.com/index.php/weblog/comments/7883/

In it, there's a suggestion that I've always thought was a sound idea: Apple should buy a distro of Linux and merge the two. Imagine, the Mac OS interface with the Linux Kernel.

Now as I've already stated, I think this is a great idea. Apple goes out and picks up a Linux distribution, maybe a Mandriva or other stable but suffering version, and cannibalize it. Take the best parts of the Linux OS and seamlessly merge them with the Mac OS. Then, offer this product in two ways:
1. As a lower-end, white-box version, that anyone could install on any regular-old x86 box, with limited support and an open-source code (just like any other Linux OS). The key here is that it's a "use-at-your-own-risk" proposition.
2. As a high-end, stable, secure and efficient Server product. This one would replace Mac OS Server, and offer all the bells and whistles a true server product should, including first-class 24/7 support (for, like, $500 a year). The source code could be open or closed, it's up to Apple. The great thing is that the product would benefit from the upgrades and updates to all the other Linux server versions.

Then at the same time, they continue to offer the original desktop Mac OS that we all know and love. This would be the higher-end product, fairly rigid in it's hardware requirements (it'll only run on Apple hardware, for instance) and it will have greater testing, stability and support. If you put it in terms of the automotive industry, the white-box version would be the Chevy Aveo and the Apple-hardware version would be the Cadillac DTS.

I'm looking for feedback on this idea, just to gauge it's viability and acceptance (and to make sure I'm not totally nutbags).

The problem with switching to a Linux kernel from Mach kernel is that it then limits Apple to what it can and cannot do with the operating system software.

If you do a little study on OS X, Apple bought NeXT to use the mach kernel, so I doubt they would want to use an open source kernel. Second, they would not need to buy a Linux Distro to use the Linux Kernel, even though, again there would have no desire to use it.

Even with that said though. Most linux source, pretty much any program, can be compiled in OS X.

The other day, I was hanging out with a friend, and he stated that OS X was the perfect mix between the Power behind linux, and a great GUI, and I completely agree.

Though your idea isn't a bad one, But OS X is already what you proposed.

The Linux kernel runs on almost any hardware.
I don't think Apple wants that. On the other hand it would be great for the end users.

Don't think it will be a very good idea. You are asking Ferrari to roll out a Honda like model

The xnu kernel is just as well or better as the horrible linux kernel. And the GUI is OSX better parts. The few parts that Apple could be interested from linuxworld they have already taken or will when they need it. And if they really wanted something that worked on all "white"-boxes, they have the knowhow and resources to do it.

http://www.kuro5hin.org/story/2009/6/26/134033/785

Excuse me while I laugh myself silly.... now, would you care to give some facts to back up your rather outrageous claim? The Linux kernel has more hardware support than _any_ other kernel available whether standalone or part of a full OS (yes, this includes Windoze. They have better driver support for desktop hardware because hardware manufacturers create their own drivers to ship with the product. Built-in driver support doesn't touch Linux). NetBSD supports more architectures (as in, x86, ARM, m68k, PPC, sparc, etc.), but Linux supports a very large number as well and has more drivers over all. Also, you'd be hard-pressed to find a benchmark (made to specifically target the _kernel_) which does not favor Linux. The closest performer I've seen is FreeBSD -- not Darwin, the OS X kernel.


To the OP: Buy a Linux distro? How do you propose they do that? You could buy the website of one, sure... but buy the code? Not possible, at least in the way you're likely thinking. Read up on the GPL (and many other open source licenses). If Apple really wanted to work with Linux, they could just create their own distro, anyways.

@QckSlvrGuyInKC: No, it wouldn't. To be sure, Apple would have to rework Linux to support their software better, but all they would have to do is send in patches to Linux for that. More likely though, since Linux is Free Software (read: open), they would maintain an external patchset, and then fork the kernel while only pulling in patches from the mainline kernel that they need. In other words, they'd create their own derivative of the kernel while letting the main kernel devs do most of the work for them, Apple devs would only need to touch (theoretically, they'd have to do a bit more than this) what they modified themselves. There isn't anything wrong with that, it's part of the flexible and open nature of Linux. It's not a product, not by itself, and you can do (almost) anything you want with it.

@jesusfreak198989: Apple's kernel _is_ open source. Darwin. NeXT was mostly concerned with what became Coaca, and such.


All the above aside, I agree that Apple should not create a Linux distro, or port OS X to Linux instead of Darwin. There is no need for it now that they have Darwin, and it would hurt their business "strategy" of locking OS X to certain PCs. Etc.

well no they dont need to go near linux but they could adopt some linux technologies like gallium3d (yea i no thats no strictly linux stuff) but what they deffo need to do is rebuild there kernel on the lates bsd kernel!

I don't see why hardware support makes qualities for a kernel, it is the inner workings which is interesting. And for Apple it is definitely not interesting, they know their hardware and have ported xnu so it runs on them. Actually the hardware support is becoming linux kernel problem, it is getting bloated as it is difficult to make one fit all source code that are easy to debug.

i agree and i also have a few solutions to the problem but for reason i can not speak about, i cant speak about any of them lol however the XNU kernel is purpose built and if the linux kernel where rebuilt to do only the things that XNU does then it would do it much better than XNU, tell me again are tousands of companies going to linux or are they going over to BSD? point proven!

Most companies goes for windows... so in that case your point proves...

While it may be other inner workings which make a kernel "interesting", driver support is the #1 thing which matters in kernel development (ask any kernel developer...). Apple can afford to be lazy here though, you are correct, because they can target their own locked-in platform specifically.

For your last statement... [citation needed]. But you haven't given citations for your previous claims, either? Anyways, I prefer an open environment where you have rapid-pace hardware innovation, which kernels that support large numbers of components (like Linux) facilitate Supporting hardware is a necessity, just as writing a scheduler as part of your kernel is a necessity. But that is a personal choice.