I've taken a few classes with some of the authors of this project and they are all amazingly talented. I remember Alex gave a presentation about a year ago on this and he mentioned how hard it was for them to wedge C++ into the Linux kernel just so they could get I/O Kit to work. It was also awesome to see the iOS Stocks app running on a Nexus 7. From what I saw it worked pretty well and aside from the uncanniness of the whole experience, it was pretty amazing to see a fully-working demo. Keep up the good work guys.
I guess the lesson here is that one shouldn't do something so awesome that one's hosting company can't withstand the HN (and likely Reddit) onslaught :-)
As an academic work this would fairly clearly fall under a fair use defense, but I wonder how the recent Oracle / Java copyright API decision would affect any potential further development / distribution / commercialization of something like this. It seems at very least it would have a huge chilling effect on anybody who might otherwise have thought to attempt it.
Even without the Oracle ruling, any attempt to commercialize something like this would get shutdown because of how it decrypts IPA files. Not to mention Apple's terms of service are very clear that iOS apps cannot be distributed outside of the App Store (the enterprise deployment "store" option being the one exception, and companies pay for that privilege and have to comply with specific guidelines around that).
It's not just the apps themselves. In its current state, the program in the paper replaces only the kernel interface and some user libraries (OpenGL); all the rest of the core Apple frameworks are copied from iOS. In theory, a Wine-like project could attempt to reimplement all of these frameworks, which is where the Oracle ruling comes in...
> In theory, a Wine-like project could attempt to reimplement all of these frameworks
You mean like the Google Ventures-backed Apportable[0]? It's already delivered 3 of the top 10 games on Android, and I'm currently using it to port a "normal" social network app from iOS to Android (using the development version which re-implements Core Animation on top of OpenGL ES 2).
And getting back to the Oracle ruling I think the fair-use defence for something like this aimed at a compatible implementation would be stronger than Google's because they were trying to fork the platform and avoid the available Java licensing offers.
Base the company in Germany, problem solved. Here it is legal to reverse engineer and reimplement stuff, as long as you're doing it to obtain compatibility with other apps/platforms.
Stated in such a general way, this is not true. It's just the case that some contractual preventions of reverse engineering might be invalid. But the moment you distribute something, patent laws can become a big problem.
Sorry about the inoperative server; many outlets have picked up on this over the last several hours, but I thought that a link to the original source would be most appropriate.
Hopefully, tomorrow morning someone will enter the server room, remove the melted plastic, and replace the previous server with a new one..
I doubt it. The jailbreak/decryption technique they used to get the IPA files was for iOS 5.1 on a 3GS. There are undoubtedly newer techniques, but actually maintaining and grabbing app files is non-trivial. And any centralized repository is instantly targeted, so you have a seedy app underground of stuff that probably isn't even current.
Then you have the fact that there is 40% overhead on Android to replicate the iOS experience. And the fact that the bluetooth and GPS components won't work.
So do you really think people will buy cheap tablets, spend time installing some hack to run older apps and then get subpar performance, just so they can have access to ios apps? Won't happen. A small sector of users might be willing to do this, but definitely not enough to disrupt anything.
I mean, this is considerably more complex than a hackintosh and we all saw how those projects just totally destroyed Apple's Mac sales. Oh, wait...
If it's anything like the hackintosh market, this could help iOS device sales if people tried and liked the iOS apps they could get on Android. Plenty of us who had a hackintosh (myself only in a virtual machine, but still) ended up switching to the Mac after trying out OS X that way.
This is amazing work. As someone who implemented a significant part of the iOS API on top of Android, I can say it's not trivial :) Disclaimer: we created an Apportable competitor for Xamarin based projects which will see the light this week.
I am more interested in people who did this than what they have achieved. The difference in capability of me and them is like distance between planets. How one can become such talented!
Much of OpenGL is implemented in hardware rather than software, and that benchmark is probably pretty close to pure OpenGL. There wouldn't be much emulation to do.
This is particularly amusing in light of the fact that there are tons of Chinese clones of the iPhone that run (often iOS-skinned) Android, e.g. the Goophone/Zophone i5. Those would make great hardware to use with this.
very cool proof of concept. adds a new, interesting dimension to multi-platform mobile development solutions. definitely noticeable lag, so you're not fooling anyone just yet with its native-ness.
how are apps loaded? Individually installed via adb? do you have access to the iTunes store on Android?
You're right, although it is systems.cs.columbia.edu specifically, so it's just a group within a department. If they're anything like my old research groups, the web server is just an old box or a Xen VM that's about to be retired, since people mostly aren't hammering the server daily.
http://en.wikipedia.org/wiki/TransGaming_Technologies#Cider