Run Mac OS X Apps On Linux?
I have the urge to commit my 24" Core 2 Duo iMac to a single Linux operating system, thus giving up the goodness of my beloved Mac OS X. I am not a stranger to Linux, but I am a stranger to running Mac apps on Linux. On my PowerPC I can use SheepShaver to run Classic apps. The Mac-on-Linux project can run OS X apps, but it requires a PowerPC, not an x86. Virtualizing and emulating are inefficient, especially given the wonderful results the WINE project has had in getting Windows apps to run on Linux. What I would like is an equivalent: a software compatibility layer that will allow Linux to run Mac OS X apps at native performance. I believe there is some additional complexity in accomplishing this. Mac OS X apps aren't just Mac OS X apps. They are Carbon. They are Cocoa. They are universal binaries. They are PPC code with Altivec. Does such a project exist yet? If not, why not?
Let's see- you take your unix kernel, build a fully rewritten, GL-optimized X Windows system... solidify your GNUStep libraries until they're fully compatible with cocoa-
Then be certain you're only running systems with SSE3 or above, since all the intel binaries are optimized for that.
Whoops- also gotta build a compatible CoreAudio system.
Oh no, looks like you're now Apple.
If you're using Mac OS X, you're already using a NeXT compatibility layer on top of a streamlined X and a really high end specialized desktop unix. The question is- why are you trying to run linux? You're using a much higher end unix system that supports X11 and has a full BSD layer, with package managers available. Why don't you just run linux apps in OS X?
Making an OS X "compatibility layer" would essentially just require you to create a shoddy set of OS X libraries- something Apple's already done better.
Take advantage of your resources.
Let me repeat this - OS X is a "mac compatibility layer" running on top of a unix kernel already- it's a totally insane waste of time to re-implement it. If you're that interested in making OS X, you should work for Apple.
Nice troll, but Wine aims for compatibility with Win32 code, not some specific version of the OS. It already supports more than 90% of the API.
"I've got more toys than Teruhisa Kitahara."
The Cocoa API is based on the OpenStep API, which has a full open specification. We've had a FOSS implementation of the OpenStep API for years: GNUstep. GNUstep has even implemented some of the Cocoa additions. Producing a Cocoa compatibility layer should be much easier than producing a Windows compatibility layer.
That's a good question. For me, my move from OS X to Linux is motivated by several reasons.
1. Portability
OSX runs on Apple Hardware. It can be hacked, sure, to run on some non-Apple systems, but widescale hardware compatibility is lacking (just look at the difficulties in getting a GeForce 8x00 series card running). With Linux, I have a much wider choice of hardware vendors.
2. Flexibility
Linux gives me a lot more control over my OS. While this is pointless for 90% of users, for me, it is very nice to be able to custom compile a kernel with the options I need, recompile any given app for the characteristics I need, and so on. I can enable, disable, theme, customize, and otherwise hack my environment to suit my needs much more easily than in OS X.
3. Security
OSX is pretty good, but I'm a little worried at Apple's highly negative response to security concerns (anyone remember the wireless driver exploit fiasco?). Also Linux offers some features that OSX doesn't currently offer, such as MAC. 10.5 is supposed to have MAC, but I can't find any info on it. While both systems are good, I perceive the Linux market as responding more credibly to security issues.
4. Software Ecosystem
OSX software is great. However, Linux offers me a great deal more choice, and that choice tends to be much more configurable to my needs. While I love the Cocoa API and what it allows apps to do, Linux is catching on fast, and I find that good GNOME or KDE apps perform most, if not all, the functionality I care about in OSX, while being beer free, and offering me more choices (i.e. Anjuta, Geany, Eclipse, KDevelop vs XCode, Eclipse). There's also a lot of software I like to use that aren't available on OSX yet. I'm starting to learn CUDA, for example, and that is not yet offered for OSX. Also, VMWare workstation offers some nice features targeted at developers (such as host-to-VM project deployment & debugging) which I haven't seen offered yet on OSX.
5. Usability
I know a lot of people disagree with this, but I find a lot of OSX's UI broken, at least for how I use my system. I find the window management to be frustrating, the Dock to be very limited, and the lack of decent app launching & management to be very frustrating (although QS does manage to solve app launching in a pretty nice way). GNOME's clean usability and massive customization have allowed me to create a Linux system that I actually find to be more usable for me than OSX.
I guess I could summarize all of that by one word: choice. For my needs and uses, I find the flexibility of Linux far more convenient.
If OSX's behaviors work well for you, and you are happy with Apple's hardware offerings, then you have the choice of OS X instead!
Isn't freedom of choice a great feeling?
multifariam.net -- yet another nerd blog
If you did this, you could run applications that just used Cocoa. Unfortunately, a lot of OS X applications also use Carbon. There is no open source version of Carbon, although a few people have written GNUstep wrappers for the parts of Carbon which are toll-free bridged with Cocoa. You would also not be able to run anything that depends on WebKit (unless you use GNUstep's SimpleWebKit as a stop-gap), QuickTime, or any of a number of other frameworks that have not yet been added to GNUstep.
WINE has had a lot of time and effort put into it because there are a huge number of proprietary Windows-only applications that would be useful on other platforms already deployed. There are far fewer essential applications for the Mac (90%+ of computer users don't use any OS X apps, after all), and so no one has bothered. For the few apps that would be worth supporting, it is probably easier to re-write them from scratch than it is to write a general purpose compatibility layer.
I am TheRaven on Soylent News