Prediction: Apple platforms will merge in 2016
I had an epiphany a few days ago while in the shower, thinking about the state of Apple operating systems, with WWDC 2015 over and the latest version of OS X 10.11 El Capitan and iOS 9 now out in public beta.
I found it funny that with OS X and iOS now having nearly simultaneous yearly releases in the fall, and OS X visibly never incrementing its major version number, something is bound to happen at WWDC next year — if only because we’re closing in on iOS 10. My first thought was whether Apple would just go ahead and call it iOS X to match the Mac side.
Actually, I think the most likely scenario given recent developments is that 2016 will be the year that iOS and OS X finally go full circle and merge back into a single software platform. The lines have been slowing blurring over the past few years, but I believe 2016 will be the year this vision finally fully realizes.
I think there are several tell-tale signs from Apple that in retrospect clearly point in this direction.
1) App Store submissions in the form of LLVM Bitcode, not just ARM binaries.
This was the biggest hint at the last WWDC that something major might be up on the hardware front.
It was very quickly mentioned in the developer State of the Union presentation, among the variety of techniques used for app “thinning”, aiming to reduce the size of app downloads by trimming down the app bundle to just the bits the user needs for their particular hardware.
The biggest change announced was that Xcode 7 would now be able to upload an intermediary binary state to Apple, nicknamed Bitcode, produced by the LLVM compiler. Then Apple’s servers will be ultimately responsible for the final phase of compiling and optimizing the app into native code appropriate for the various CPU platforms (32-bit, 64-bit, various ARM architectures, etc). Think something relatively similar to Java VM byte code, except at a much lower level.
Apple even said that all iOS apps will soon be required to use this new Bitcode delivery system when submitting their apps to the App Store for review. It is actually required for watchOS 2 apps, remains optional (but encouraged) for iOS 9, but will surely become mandatory after this.
My first thought when I heard about Bitcode was that it was a ploy to allow Apple to switch to a new CPU platform of their own — we might conceivably see the end of ARM on iOS very soon, and this way Apple would be able to automatically “upgrade” all Bitcode-submitted apps magically for the new devices. This would be a much cleaner and faster way to handle a major hardware transition like when the Mac transitioned to Intel from PowerPC a few years back. In their presentation, the Apple engineer even refered to Bitcode as a way to “future-proof” apps for new hardware.
Another possible application of such a technology? Overnight, Apple could decide they want to support the App Store on a system not unlike the iOS simulator currently included with Xcode. In theory, with this system in place, it would be a relatively minor change to the simulator to give it an Apple Store app that would be able to purchase and download iOS apps and get full-fledged Intel iOS binaries ready to run on a Mac.
Pretty much all iOS frameworks and technologies are already available as Intel versions for the simulator. Developers like myself already use the simulator all the time to develop the apps before sending them to Apple, so this would effectively extend this technology to become a mainstream application.
2) The Mac and iOS developer programs were just merged into a single Apple Developer Program.
As the overlap between the OS X and iOS platforms becomes larger, it makes sense to simplify things for the developers and consolidate the developer programs. As of WWDC 2015, there is now just one unified developer program that gives access to all resources for both the Mac and iOS platforms, for a single $99/year fee.
While the disappearance of the free Safari developer program irked a number of developers, the bigger picture here is that this points to the completion of the internal restructuring of design teams at Apple, led by Jony Ive a couple of years ago.
This was hinted when Apple consolidated their teams along software and hardware lines, instead of along OS platforms, and developers are starting to feel the end result of this transition.
3) The new iOS 9 multitasking support opens the door for Mac-like iOS apps.
iOS 9 will for the first time allow several apps to run concurrently on the same iPad screen, via features named Split View and Slide Over. No doubt future iPad models will be even more capable of this feat, with increased RAM and CPU power. At the same time, the new OS X El Capitan showed new ways for Mac apps to use Split View while in the full-screen mode now available to apps, not unlike what you will be getting on the iPad with iOS 9.
With these latest changes, iOS 9 apps could conceivably run in their own window on a Mac desktop, even going as far as being resized dynamically.
While this is arguably a new public-facing feature, Apple has also been leading developers to make device-agnostic apps using technologies such as Auto Layout, now available for both iOS and OS X. In 2014, Xcode 6 even introduced a resizable iOS simulator so that apps could be tested among virtual devices of any dimensions. This was a strong hint that the iPhone 6 and 6 Plus would come out a few months later with their larger screens.
With Xcode 7, there is renewed emphasis on this aspect as apps now have to be able to respond to size changes dynamically as they may be sharing the screen with another app. This concept would make such resizable apps fit right at home as a window on OS X.
This convergence of windowing concepts, paired with the strong push for developers to develop device-agnostic interfaces in their iOS apps, is a strong hint that iOS apps will be more and more comfortable to fit a Mac desktop windowing system.
As a side note, it is interesting to see that Apple is also uniting all system fonts around their new San Francisco font face. As a result, Mac and iOS apps will look and feel more like each other than ever before.
4) Metal superseding OpenGL on both iOS and OS X.
Besides the huge performance gain that the renewed importance of the Metal framework at the system level will bring to all applications, this points to Apple’s renewed focus on “close-to-the-metal” GPU utilization. Again, this will work better if Apple also controls the design of the hardware that Metal talks to.
Metal on the Mac in OS X El Capitan will have to be optimized for current-generation GPUs from AMD, nVIDIA and the like. These GPUs have their own architectures and shading languages, well within the control of their original manufacturers. In contrast, the reason Metal was first available on iOS probably has to do with Apple’s tight control over the chips included on that platform — their very own A* series.
History tells us that Apple relishes total control over their hardware chain — it should be obvious that they can’t wait to ditch these third-party GPU/CPU manufacturers the minute they are able to. Cross-platform Metal is their Trojan horse to do just that.
5) The new Macbook is a glimpse at Apple’s next-generation Mac hardware.
The last generation of MacBook was arguably a major change, and it put it more in line with the iPad than with the Mac. A lot of the attention it got was because of the sole USB-C port it is equipped with, but when looking at the device its thinness and even reduced performance really makes it more in the class of a current-generation iPad.
The Force Touch trackpad also mirrors technology first introduced with the Apple Watch, and sure to further extend to more Macs and the next generation of iOS devices coming out in the fall.
Soon the most glaring difference between an iPad and a MacBook will be the presence of a physical keyboard attached to the screen. Conversely, I wouldn’t be surprised to see some form of “MacBook Touch” soon. I bet that when the new unified Apple OS is announced next year, this new model of Macbook will look prescient in retrospect.
It’s actually interesting that Microsoft themselves tried to do this laptop-mobile unification with the Windows 8 fiasco on Surface tablets a few years ago, to very underwhelming results. I bet once again Apple will show us how to do this kind of transition right.
With these facts in mind, I would like to put my thoughts out on the record and wager a bet. This post might end up being a nice bowlful of claim chowder come next year, but I am going to make a number of specific predictions as of this July 2015.
My bets for the near future of Apple platforms are the following :
- A unified Apple OS will be announced at WWDC 2016. Its name is not really important, whether they’ll call it OS X, iOS X, or something entirely new. I actually wouldn’t be surprised if it’s simply called “Apple OS.” The point being that it will be the same OS platform running on both the Mac and iPhone/iPad. Macs will be able to download and run iOS apps natively alongside legacy Mac apps. Legacy Intel Macs will be able to run current iOS apps natively through transparent Bitcode recompilation happening in the App Store. I don’t believe the reverse will be true, however — no legacy Mac apps on iPad.
- The two remaining major Apple software platforms will then be the new unified “Apple OS” along with watchOS (likely version 3 at that point). The new unified Apple OS will be focused on bridging the last remaining gaps between Mac and iOS devices.
- Hardware platforms will also merge. We will start seeing new Macs equipped with Apple-designed high-performance “AX” 64-bit CPUs, the same line that will be used in the next generation of Apple mobile devices. These CPUs might be ARM-based, or will retain some kind of compatibility with the Intel instruction set, or they even be something entirely new, but it will not matter that much. Apple will stop producing new Mac hardware with major third-party components designed by Intel, AMD or nVIDIA — possibly as far as the Mac Pro. Force Touch will become mainstream on all new devices.
- Major hardware rebranding across the lines. The Mac brand name might even change for the first time ever, as the lines between iPad and MacBook will continue to blur. Hell, considering the “i” prefix has been mostly phasing out lately, I wouldn’t be surprised if we were going to have simply an “Apple Phone.” Apple Pad? Apple Computer?
At this point, I believe these are the next logical steps. Certainly Apple has been preparing for this moment for a long time already, and it seems like the day of reckoning is at hand.
Apple surprised everybody when they were the first to introduce a 64-bit “desktop-class” A7 processor with the iPhone 5s, taking the entire mobile industry by surprise. I think they will repeat the surprise by producing their own truly groundbreaking desktop-grade CPUs for the next generation of Macs and mobile devices, rendering their dependence on Intel technologies effectively obsolete.
This will be a major sea change in the computing world, sure to be followed by cries from detractors continually whining that Apple is not able to innovate, even as trillion-dollar AAPL will become the first computing company with full end-to-end control over the entire hardware and software in their products. Mark my words.
