You're forgetting backward compatibility, which is important.
Based on that statement Apple should never have moved from Motorola, never have moved from PPC, and never should have stripped macOS down and built it up to make iOS (on ARM), and never should have adopted 64-bit on macOS or iOS. Yet, in all those cases they not only succeeded but have continued to make the transitions even smoother by building their OSes and developer tools that make easy work for both Apple's in-house apps and 3rd-party apps.
What you're ignoring is that with both of those transitions, Apple had a strategy for backwards compatibility, because in both cases, the architecture they were switching to was sufficiently powerful to emulate the old one and keep existing software running with decent performance. That is the difference here. If ARM ever pulls ahead of Intel such that x86 could be decently emulated, then a transition along those lines could be done. We're nowhere near that point, though.
But, hey, let's ignore Apple's history, let's ignore that Apple's notebook and desktop user base is considerably higher now, let's ignore advancements like App Thinning and Bitcode, and let's ignore that Intel is plateauing in their chip performance and power efficiency, let's ignore that Intel's roadmap is hindering Apple's Mac sales, let's ignore Intel's chip costs are keeping Apple from accessing millions more Mac users at a less-expensive end of the PC market, and let's ignore that Apple's A-series chips are already outperforming Intel's Core-M processors (used in the MacBook) while using only 60% of its power.
Let's also ignore that Core-M is pretty close to the absolute low end of Intel's lineup, and ARM doesn't have anything that can compete with any of the rest of it.
Are we really going to be pedantic enough to bring clearly irrelevant things like watchOS and tvOS into this? I think it's pretty clear what I meant.
I see nothing irrelevant about these OSes or their associated products. On top of that, Apple has at least one OS for the iPod, I think based on Pico OS, that it's used for 15(?) years now, and then you have their firmware/OS for their AirPort devices and all the ARM-based chips that run innumerable other systems. We've been told that the T1 chip is based off of the S-series chip in Watch, but we don't know if it's running a stripped down version of watchOS—not unlike how they stripped down macOS and then built it up to be iOS—or if it's more inline as a basic firmware that runs the Touch Bar, Touch ID, and Apple Pay. We don't even know if the T1 has a GPU for the Touch Bar's display or if that is pulled from the primary system GPU, or if macOS runs that UI or if it's handled by the T1 which then relates all the data seamlessly to MacOS. Regardless, Apple is very deep into ARM development with many OSes and firmwares separated by varying degrees of complexity.
You're forgetting backward compatibility, which is important.
Based on that statement Apple should never have moved from Motorola, never have moved from PPC, and never should have stripped macOS down and built it up to make iOS (on ARM), and never should have adopted 64-bit on macOS or iOS. Yet, in all those cases they not only succeeded but have continued to make the transitions even smoother by building their OSes and developer tools that make easy work for both Apple's in-house apps and 3rd-party apps.
What you're ignoring is that with both of those transitions, Apple had a strategy for backwards compatibility, because in both cases, the architecture they were switching to was sufficiently powerful to emulate the old one and keep existing software running with decent performance. That is the difference here. If ARM ever pulls ahead of Intel such that x86 could be decently emulated, then a transition along those lines could be done. We're nowhere near that point, though.
But, hey, let's ignore Apple's history, let's ignore that Apple's notebook and desktop user base is considerably higher now, let's ignore advancements like App Thinning and Bitcode, and let's ignore that Intel is plateauing in their chip performance and power efficiency, let's ignore that Intel's roadmap is hindering Apple's Mac sales, let's ignore Intel's chip costs are keeping Apple from accessing millions more Mac users at a less-expensive end of the PC market, and let's ignore that Apple's A-series chips are already outperforming Intel's Core-M processors (used in the MacBook) while using only 60% of its power.
Let's also ignore that Core-M is pretty close to the absolute low end of Intel's lineup, and ARM doesn't have anything that can compete with any of the rest of it.
1) You're assuming that Apple has to emulate anything. We now have a Mac App Store and Xcode tools that have improved so much that when this happens I don't think we'll see another Rosetta.
2) You can't ignore the 12" MacBook when it's being used by Apple right now.
3) You're making an erroneous assumption that if Apple adds a ARM-based Mac that you have to look it as being able to replace the fastest Mac in the lineup instead of looking it at as being a low-cost, low-power machine that will allow Apple to sell millions of additional Macs that will have millions of available apps on the App Store. That's where this will be heading, not to the Mac Pro and not to the MacBook Pro, at least not years down the road. For me, a sub-$800 notebook running ARM with speeds that far exceed the current 12" MacBook is not something I'd ever want. You need to stop thinking about what's best for you and instead think about what is best for Apple and other customers.
The difference between iOS and macOS (and tvOS and watchOS) is NOT what processor architecture they run on, but how we interact with them. "Adding a mouse pointer to iOS" is not a trivial thing; it requires pretty much exactly the kind of redesign and rethinking that would turn it into...macOS.
Well, the question you need to ask yourself is; what do most people need a full-fledged laptop for? What's the use case? What does your average user, say a person who writes papers for work/school/etc, need in a laptop that s/he doesn't get from a tablet?
Apple's strategy for the iPad has seemingly been to close the gap between tablets and full-fledged computers for some time now, and make tablets the new default for consumers in the "post-PC" future (see: the Jobs cars/trucks analogy). The iPad Pro is clearly a step in this direction; the built-in keyboard positions the iPad in a market where you'd typically have needed a laptop; typing up long documents. Did this fundamentally change the nature of the iPad, or "turn it into macOS"? Not really; it was an additive feature. The touch input is still the default, but with an optional Bluetooth keyboard, or the Apple flip case, you have an alternative input method that's useful for a lot of tasks.
So the question now becomes, what's still missing, that you'd need for day-to-day computing? And the answer is: not a whole lot, actually! A tablet, a keyboard, and some sort of stand can replicate most of what most non-techie people use a computer for, in fact. The one hole in it, honestly, is revising text, which, at least if you have large hands, can make selecting individual words and/or letters difficult to do with precision on a touch screen. That's what you introduce a trackpad for: for making fine corrections. It can also be used to simulate taps on the touch screen, yes, but that's not the primary purpose of it; it's not to replace the touch screen, it's to do the small percentage of tasks which the touch screen is not suited for. This, like the keyboard, is an additive feature. It's not hard to add, it doesn't change the UI in any appreciable way, and if you don't get a flipcase with a trackpad on it, there's no change for you at all. It also doesn't turn iOS into macOS; after all, Android and Windows Phone both have mouse support, and they're not macOS.
In summary, this is 1) a lot easier to do, 2) introduces a lot fewer problems and headaches, and 3) actually meshes with Apple's vision. When it actually happens, you'll all think it's the best thing ever, I'm sure.
Are we really going to be pedantic enough to bring clearly irrelevant things like watchOS and tvOS into this? I think it's pretty clear what I meant.
I see nothing irrelevant about these OSes or their associated products. On top of that, Apple has at least one OS for the iPod, I think based on Pico OS, that it's used for 15(?) years now, and then you have their firmware/OS for their AirPort devices and all the ARM-based chips that run innumerable other systems. We've been told that the T1 chip is based off of the S-series chip in Watch, but we don't know if it's running a stripped down version of watchOS—not unlike how they stripped down macOS and then built it up to be iOS—or if it's more inline as a basic firmware that runs the Touch Bar, Touch ID, and Apple Pay. We don't even know if the T1 has a GPU for the Touch Bar's display or if that is pulled from the primary system GPU, or if macOS runs that UI or if it's handled by the T1 which then relates all the data seamlessly to MacOS. Regardless, Apple is very deep into ARM development with many OSes and firmwares separated by varying degrees of complexity.
Sure, sure, let's build a laptop replacement on watchOS. Makes sense.
You're forgetting backward compatibility, which is important.
Based on that statement Apple should never have moved from Motorola, never have moved from PPC, and never should have stripped macOS down and built it up to make iOS (on ARM), and never should have adopted 64-bit on macOS or iOS. Yet, in all those cases they not only succeeded but have continued to make the transitions even smoother by building their OSes and developer tools that make easy work for both Apple's in-house apps and 3rd-party apps.
What you're ignoring is that with both of those transitions, Apple had a strategy for backwards compatibility, because in both cases, the architecture they were switching to was sufficiently powerful to emulate the old one and keep existing software running with decent performance. That is the difference here. If ARM ever pulls ahead of Intel such that x86 could be decently emulated, then a transition along those lines could be done. We're nowhere near that point, though.
But, hey, let's ignore Apple's history, let's ignore that Apple's notebook and desktop user base is considerably higher now, let's ignore advancements like App Thinning and Bitcode, and let's ignore that Intel is plateauing in their chip performance and power efficiency, let's ignore that Intel's roadmap is hindering Apple's Mac sales, let's ignore Intel's chip costs are keeping Apple from accessing millions more Mac users at a less-expensive end of the PC market, and let's ignore that Apple's A-series chips are already outperforming Intel's Core-M processors (used in the MacBook) while using only 60% of its power.
Let's also ignore that Core-M is pretty close to the absolute low end of Intel's lineup, and ARM doesn't have anything that can compete with any of the rest of it.
1) You're assuming that Apple has to emulate anything. We now have a Mac App Store and Xcode tools that have improved so much that when this happens I don't think we'll see another Rosetta.
Not all software people are using comes from the Mac App Store. Quite the opposite, actually; many of the most prominent apps people use are not on the MAS. Microsoft Office, Photoshop, etc.
The bitcode concept is also currently being used only for iOS (and watchOS and tvOS, which I am ignoring in the context of this discussion). For macOS, there are additional problems with it that you don't have with iOS; for example, apps can link against native third-party libraries on macOS, which isn't possible with the bitcode concept. You'd need some sort of emulation.
2) You can't ignore the 12" MacBook when it's being used by Apple right now.
3) You're making an erroneous assumption that if Apple adds a ARM-based Mac that you have to look it as being able to replace the fastest Mac in the lineup instead of looking it at as being a low-cost, low-power machine that will allow Apple to sell millions of additional Macs that will have millions of available apps on the App Store. That's where this will be heading, not to the Mac Pro and not to the MacBook Pro, at least not years down the road. For me, a sub-$800 notebook running ARM with speeds that far exceed the current 12" MacBook is not something I'd ever want. You need to stop thinking about what's best for you and instead think about what is best for Apple and other customers.
You really want to split the lineup among two architectures with differing software compatibility attributes? In addition to causing massive headaches for developers (gotta test everything on both architectures!), that's basically the Surface/Surface Pro strategy, which blew up in Microsoft's face. You'll have "this app can run on this current Mac, but not that current Mac, and this other app can only run on the other Mac but not the first one." When Microsoft did it, it confused the heck out of customers and caused a great deal of backlash, and Microsoft ended up moving the whole line back onto Intel. And Microsoft actually had cross-architecture apps, with the CLR (which ran in a VM, similar to Java). It still blew up on them. I don't think that's worth saving a little bit of battery life. Go all the way, or go home.
Are we really going to be pedantic enough to bring clearly irrelevant things like watchOS and tvOS into this? I think it's pretty clear what I meant.
I see nothing irrelevant about these OSes or their associated products. On top of that, Apple has at least one OS for the iPod, I think based on Pico OS, that it's used for 15(?) years now, and then you have their firmware/OS for their AirPort devices and all the ARM-based chips that run innumerable other systems. We've been told that the T1 chip is based off of the S-series chip in Watch, but we don't know if it's running a stripped down version of watchOS—not unlike how they stripped down macOS and then built it up to be iOS—or if it's more inline as a basic firmware that runs the Touch Bar, Touch ID, and Apple Pay. We don't even know if the T1 has a GPU for the Touch Bar's display or if that is pulled from the primary system GPU, or if macOS runs that UI or if it's handled by the T1 which then relates all the data seamlessly to MacOS. Regardless, Apple is very deep into ARM development with many OSes and firmwares separated by varying degrees of complexity.
Sure, sure, let's build a laptop replacement on watchOS. Makes sense.
That's what's you read when I mentioned that the T-series chip is based on the S-series chip? Jesus fucking christ!
durandal_1707 said: You'd need some sort of emulation.
No.
Aha! Well, then.
You really want to split the lineup among two architectures with differing software compatibility attributes?
I don't want to do anything. This is about what makes sense for Apple in the longterm.
In addition to causing massive headaches for developers (gotta test everything on both architectures!)
Gee, and that's been a huge failure for Apple the dozen times they've had to adopt their architecture¡
Apple has never had two "current" architectures for the macOS lineup at the same time. It's always been a transition from one architecture to the next one, followed by dropping the old one entirely.
that's basically the Surface/Surface Pro strategy
No it's not.
Well, okay then. In the face of that brilliant rebuttal, I think I'm done here.
Apple has never had two "current" architectures for the macOS lineup at the same time. It's always been a transition from one architecture to the next one, followed by dropping the old one entirely.
Yes, they have. And that's just taking about transitioning their current selling HW, not having support for both architectures at the same time once the transition was complete. THIS IS EASILY RESEARCHABLE STUFF!!!
Apple has never had two "current" architectures for the macOS lineup at the same time. It's always been a transition from one architecture to the next one, followed by dropping the old one entirely.
Yes, they have. And that's just taking about transitioning their current selling HW, not having support for both architectures at the same time once the transition was complete. THIS IS EASILY RESEARCHABLE STUFF!!!
Sigh. I know I shouldn't reply, but I will, just this once. This will be my final post in this thread, however.
Yes, this is easily researchable stuff. Furthermore, I was alive and heavily into Mac stuff at the time all these transitions occurred (yes, I'm old). These are Apple's major Mac architecture transitions:
68k -> PowerPC: Apple released its first Power Macs at the end of 1994. In 1995, of the 47 Mac models released that year, all but 7 were based on the PowerPC chip. Of the stragglers, three were low-end PowerBooks for which Apple hadn't yet managed to make a mobile PowerPC cheaply enough, three more were the edu-focused LC 580 and its Performa variants, none of which were sold in the United States, and the last one was the Performa 640, a minor update to the 630, kept around for its ability to dual-boot Windows (via an onboard Intel processor) for those who needed that feature. Every new Mac model introduced in 1996 was PowerPC-based. At no point during the transition was it unclear that 68k was a legacy architecture on its way out, and PowerPC was the future.
PowerPC -> Intel: Apple announced in 2005 that they would be transitioning. In 2006, every new Mac model introduced had an Intel processor. By August 2006, Apple announced "Transition complete", as all of their lineup was now Intel-based. At no point were we under the impression that PowerPC had any future with Apple during this period.
That's how Apple does architecture transitions. Get onto the new one, then get off the old one. There's a clear progress from and to, and during the whole process one architecture will be "current," and the other will be "legacy." At no point has Apple introduced a totally different processor architecture only for one segment of their Mac lineup and then continued expending resources to actively develop the ecosystems for both architectures in parallel. That's a tremendous waste of resources, and is not how Apple operates. What Apple does is maintain the old one for a short transitionary period, and then drop it.
And no, before you say something, I'm not counting different software-compatible chip families within the same architecture, like G4 vs. G5 (which itself was only a result of IBM's failure to make a laptop version of the G5, the frustration over which was a major motivator for the Intel transition in the first place). If you want, you could count 32-bit Intel vs. 64-bit Intel, since that did require a developer transition, but even then, by 2007, there were no new 32-bit machines released by Apple.
In summary: ARM is not ready to replace Intel in the Mac line, and Apple has no reason to needlessly complicate the Mac lineup when their existing strategy of making the iPad into a desktop replacement not only can go much more smoothly, but also will probably make them more money in the long run anyway.
However, among the operating systems that Apple owns, iOS is the one with the existing ARM-based architecture and software library.
Again, you're wrong.
What part of that statement is wrong, that iOS's existing infrastructure and its app library run on ARM or that macOS and its apps don't?
1) You have no idea what architecture Apple has for macOS yet you made an absolute statement without considering Apple's long history of building their OSes to work with other archtiectures in-house.
Well, I was referring to actual, working, shipping code, not secret prototypes, but aside from that, you're leaving out the "and" in that sentence. Where is the existing software library for macOS on ARM?
2) You made another absolute statement that iOS is "the one" ARM-based OS they have when they have many that are in shipping products right now.
Are we really going to be pedantic enough to bring clearly irrelevant things like watchOS and tvOS into this? I think it's pretty clear what I meant.
It was important to point out because it highlights the common misunderstanding that the difference between macOS and iOS is the cpu architecture it runs on.
In summary: ARM is not ready to replace Intel in the Mac line, and Apple has no reason to needlessly complicate the Mac lineup when their existing strategy of making the iPad into a desktop replacement not only can go much more smoothly, but also will probably make them more money in the long run anyway.
Yes, Apple has SHIPPING iDevices that already outperform the 12" MBP and 2013 MBPs and all with a much high power efficiency. Your repeated problem is you have this incredibly ridiculous notion in your head that if a shipping iPhone's performance doesn't far exceed that of the Mac Pro or iMac that there's no R&D that Apple could be working on to make ARM viable for macOS. That's both ridiculous and myopic.
Comments
Let's also ignore that Core-M is pretty close to the absolute low end of Intel's lineup, and ARM doesn't have anything that can compete with any of the rest of it.
2) You can't ignore the 12" MacBook when it's being used by Apple right now.
3) You're making an erroneous assumption that if Apple adds a ARM-based Mac that you have to look it as being able to replace the fastest Mac in the lineup instead of looking it at as being a low-cost, low-power machine that will allow Apple to sell millions of additional Macs that will have millions of available apps on the App Store. That's where this will be heading, not to the Mac Pro and not to the MacBook Pro, at least not years down the road. For me, a sub-$800 notebook running ARM with speeds that far exceed the current 12" MacBook is not something I'd ever want. You need to stop thinking about what's best for you and instead think about what is best for Apple and other customers.
Apple's strategy for the iPad has seemingly been to close the gap between tablets and full-fledged computers for some time now, and make tablets the new default for consumers in the "post-PC" future (see: the Jobs cars/trucks analogy). The iPad Pro is clearly a step in this direction; the built-in keyboard positions the iPad in a market where you'd typically have needed a laptop; typing up long documents. Did this fundamentally change the nature of the iPad, or "turn it into macOS"? Not really; it was an additive feature. The touch input is still the default, but with an optional Bluetooth keyboard, or the Apple flip case, you have an alternative input method that's useful for a lot of tasks.
So the question now becomes, what's still missing, that you'd need for day-to-day computing? And the answer is: not a whole lot, actually! A tablet, a keyboard, and some sort of stand can replicate most of what most non-techie people use a computer for, in fact. The one hole in it, honestly, is revising text, which, at least if you have large hands, can make selecting individual words and/or letters difficult to do with precision on a touch screen. That's what you introduce a trackpad for: for making fine corrections. It can also be used to simulate taps on the touch screen, yes, but that's not the primary purpose of it; it's not to replace the touch screen, it's to do the small percentage of tasks which the touch screen is not suited for. This, like the keyboard, is an additive feature. It's not hard to add, it doesn't change the UI in any appreciable way, and if you don't get a flipcase with a trackpad on it, there's no change for you at all. It also doesn't turn iOS into macOS; after all, Android and Windows Phone both have mouse support, and they're not macOS.
In summary, this is 1) a lot easier to do, 2) introduces a lot fewer problems and headaches, and 3) actually meshes with Apple's vision. When it actually happens, you'll all think it's the best thing ever, I'm sure.
The bitcode concept is also currently being used only for iOS (and watchOS and tvOS, which I am ignoring in the context of this discussion). For macOS, there are additional problems with it that you don't have with iOS; for example, apps can link against native third-party libraries on macOS, which isn't possible with the bitcode concept. You'd need some sort of emulation.
You really want to split the lineup among two architectures with differing software compatibility attributes? In addition to causing massive headaches for developers (gotta test everything on both architectures!), that's basically the Surface/Surface Pro strategy, which blew up in Microsoft's face. You'll have "this app can run on this current Mac, but not that current Mac, and this other app can only run on the other Mac but not the first one." When Microsoft did it, it confused the heck out of customers and caused a great deal of backlash, and Microsoft ended up moving the whole line back onto Intel. And Microsoft actually had cross-architecture apps, with the CLR (which ran in a VM, similar to Java). It still blew up on them. I don't think that's worth saving a little bit of battery life. Go all the way, or go home.
I don't want to do anything. This is about what makes sense for Apple in the longterm.
Gee, and that's been a huge failure for Apple the dozen times they've had to adopt their architecture¡
No it's not.
That's what's you read when I mentioned that the T-series chip is based on the S-series chip? Jesus fucking christ!
Apple has never had two "current" architectures for the macOS lineup at the same time. It's always been a transition from one architecture to the next one, followed by dropping the old one entirely.
Well, okay then. In the face of that brilliant rebuttal, I think I'm done here.
Yes, this is easily researchable stuff. Furthermore, I was alive and heavily into Mac stuff at the time all these transitions occurred (yes, I'm old). These are Apple's major Mac architecture transitions:
68k -> PowerPC: Apple released its first Power Macs at the end of 1994. In 1995, of the 47 Mac models released that year, all but 7 were based on the PowerPC chip. Of the stragglers, three were low-end PowerBooks for which Apple hadn't yet managed to make a mobile PowerPC cheaply enough, three more were the edu-focused LC 580 and its Performa variants, none of which were sold in the United States, and the last one was the Performa 640, a minor update to the 630, kept around for its ability to dual-boot Windows (via an onboard Intel processor) for those who needed that feature. Every new Mac model introduced in 1996 was PowerPC-based. At no point during the transition was it unclear that 68k was a legacy architecture on its way out, and PowerPC was the future.
PowerPC -> Intel: Apple announced in 2005 that they would be transitioning. In 2006, every new Mac model introduced had an Intel processor. By August 2006, Apple announced "Transition complete", as all of their lineup was now Intel-based. At no point were we under the impression that PowerPC had any future with Apple during this period.
That's how Apple does architecture transitions. Get onto the new one, then get off the old one. There's a clear progress from and to, and during the whole process one architecture will be "current," and the other will be "legacy." At no point has Apple introduced a totally different processor architecture only for one segment of their Mac lineup and then continued expending resources to actively develop the ecosystems for both architectures in parallel. That's a tremendous waste of resources, and is not how Apple operates. What Apple does is maintain the old one for a short transitionary period, and then drop it.
And no, before you say something, I'm not counting different software-compatible chip families within the same architecture, like G4 vs. G5 (which itself was only a result of IBM's failure to make a laptop version of the G5, the frustration over which was a major motivator for the Intel transition in the first place). If you want, you could count 32-bit Intel vs. 64-bit Intel, since that did require a developer transition, but even then, by 2007, there were no new 32-bit machines released by Apple.
In summary: ARM is not ready to replace Intel in the Mac line, and Apple has no reason to needlessly complicate the Mac lineup when their existing strategy of making the iPad into a desktop replacement not only can go much more smoothly, but also will probably make them more money in the long run anyway.