1. The fact of the matter is that BOTH Microsoft and Apple are being cagey on the subject. That may be the absence of evidence but it is not the evidence of absence.
2. If you flash back to WWDC 2005 and the time period leading up to and even past Macworld 2006 when the first Intel iMac and 15" MacBook Pro were announced, Apple's stance on getting Windows to boot on their systems was that they were not going to do anything to expressly allow it. It wasn't until April of that year when they'd issue firmware updates for those two models and the then-recently announced Intel Mac mini that allowed for booting and installation of Windows XP along with creation of a drivers disc via Boot Camp.
Given both of these things, fine, Craig said in the interview that there will be no direct booting of any other operating system. If you believe that is set in stone, then you really don't know your Apple history.
What the context of a situation in history matters. Espectially if trying to "reuse" it as a template to use for something in the present. There are large amount of mismatch here which makes blindly mapping somewhat dubious to put into the "inevitible given history" category.
First, the architecture being jumped into isn't clearly dominated by Windows. Apple moving into x86 where Windows was the overwhelmingly dominnate player meant that yeah at some point Apple did have to go on bended knee and kiss the Windows ring. Apple did NOT want to within 50ft of BIOS ( was on Openfirmware before and IBM/Apple hoped to move the x86 world forward on EFI. Didn't happen) Inertia, Windows, and some system players put a giant cork in that. UEFI ended up initially being a huge gift of greatly extended lifetime for BIOS. Apple got stuck with EFI with BIOS bolted on for a long while.
If there is a "history" lesson for Apple here then it is more so "don't get stuck with legacy boot solution because they are tarpits. " If there is no dominate secure boot player then Apple can blaze their own trial if they move before calcification sets in.
At the moment there are 3-4 Windows 10 Arm systems out there. But they collectively probably don't outnumber of the number of iPad Pros out there ( let alone general iPads ). If there is a player here who has done the "mostest" in terms of setting the market leading position it is Apple, not Microsoft. ( If Microsoft hadn't 'failed' at phones then maybe, but they did so 2020 isn't 2006 in the ARM space. Not even close. ). If Apple releases a laptop in the same price range as Microsoft's Surface Pro X then within 2-4 months it will be Apple who has the dominate numbers on ARM laptops > $1K.
Second, Windows 10 on Arm is mired in win32 mud. Apple has a x86_64 legacy app solution that mostly works. Windows doesn't. The latter is keeping more folks on Windows 10 x86_64 ( in fact some are still clinging to x86 32-bit. Microsoft is only now getting to put to start to restriction 32-bit version of the OS to OEMs. ) . The other problem is so far have been betting the ARM solutions on Qualcomm and that too is a bit mired. At least in terms of performance.
Later this year will bring slower than 8cx systems to Windows 10 ARM
www.anandtech.com
We'll see how that goes. ( it probably will work better in emerging markets where Apple is much weaker and system cost more critical , but only Apple putting higher still prices on their own systems is going to keep them from blowing those out of the water in competition against those. Pershaps there is a Cortex-X that Microsoft is going to push out the door quick but again strategically they are not in the "you much come and kiss the ring" position at all.
Third, hardware virtualization is an order of magnitude better than what virtualzation was in 2006. It is 14 years later folks... things have dramatically changed. For non GPU intensive applications the different between "raw metal" Windows ARM app "foo" and virtualized ARM cpu app "foo" is hardly anything. All the user space ( non kernel ) code excecution is exactly the same. There only place loose ground is on the kernel calls. There barely even any kernel call virtualzation support in 2006. And the current 2020 implementations blow the doors off of that initial , "version 0.1" stuff.
The big gap now is pragmatically more so in virtual GPUs versus paravirtiualized GPUs. If can give a guest OS just as lightweight overhead interface to the GPU as the CPU then the "need" for bare metal OS hosting dramatically drops away. In 2006 there were no web streaming gaming services with user hosted on paravirtualized GPUs. In 2020 there is. Apple does need to do very substantive work to make that happen. There are major pieces done. The Apple Silicon IOMMU mapping can be done per device ( and a GPU is a device). If Apple added a guest OS mapping down to a vGPU interface on the GPU hardware the performance loss overhead would be much, much smaller. Adding hardware support for virtual GPUs would only grease the wheels even more.
If Apple doesn't want to do the paravirtualization work then yeah they would have painted themselves into a corner where they may "have to" do the boot level stuff. This is a substantively different point in history though because they do have another option.
Fourth, the major big change is that Microsoft has generally moved away from charging money for the OS. They have shifted more to a bundled with the system model. To throw further "gas" on that 'fire' they have jumped into the systems selling business themselves ( in a limited way ). So Microsoft probably won't the MacBooks to 'define' Windows 10. First, Apple isn't going to bundle the OS with the system. So there is the huge conflict with the current Microsoft model on getting paid for the OS. Second, Microsoft rents virtualiizated Windows ( again another revenue flow that doesn't need Macs to enable it at all). Same paravirtualizatin of GPUs again for hypervisors of Windows for Guest OS. That is a way around many of the "bare metal" requirements requests. If Microsoft solves that problem on their own business they can grow Windows 10 ARM there also with reoccuring revenue. ( 2020 revenue models not 2006 revenue models )
In summary, there are number of issues here where Apple and Microsoft can butt heads on what to do here for a substantial amount of time. If Microsoft really, really, really needs that bundled system revenue then there is little reason for them to "loose money' than enable Apple. Similarly Apple doesn't have to "kiss the ring'" since no one has already won "classic form factor" PC ARM space yet.
The Boot RAW on ARM does nothing for the folks grumbling about need raw x86_64 binary execution to match what deployed in cloud or uber, super , duper 3D graphics mode with princess and the pea latency requirements.
32-bit x86 WIN32 apps do JUST WORK in Windows 10 on ARM64. This is not news. And, while Windows 10 on ARM64 isn't widely out there, the reason is that it hasn't been well received DUE TO POOR HARDWARE SPEEDS. If Microsoft wants an opportunity to prove that Windows 10 on ARM64 can be worth a crap, THIS IS THAT GOLDEN OPPORTUNITY.
Without access to "big" GPUs for Windows x86_64 apps there is a big gap there in terms of the performance requirements for the 'hard core' Besides there is not going to be a "one for one" growth of Windows 10 ARM with Mac sales because Win10 won't be bundled with the system sales. You are only talking a small fraction. That won't solve their problem.
Microsoft's root cause problem is either need someone besides Qualcomm to jump into Cortex-X1 quicker or someone else. Technically if purely want raw performance numbers this
www.anandtech.com
could work. The ARM server CPUs will trickle down into Windows 10 space and those will probably better top end performance enablers than Apple Silicon will be for a somewhat long time. There is tons of action in the ARM CPU space to cover than just Apple for Microsoft to be putting work into. If peephole down to lightweight laptops then yeah Apple is more of a player. Outside of that scope though, they aren't the "only game in town".
Actually, no, the louder you complain, the more they listen. Case in point, the butterfly keyboard. Case in point the demand that precipitated the original release of Boot Camp in 2006 against similar notions of "it's never going to happen, don't even ask". Attitude like the one you're espousing is what gets people nowhere .
That is more forum groupthink induced delusion than accurate.
First, Apple moved on from butterfly keyboards after over a year of covering all repairs. It is highly likely that the money was a bigger "pain" point than the "loudness" of user forms and feedback submissions. ( service support queue time consumed, keyboard change request ... which is a
LARGE unit swap out since other stuff baked into the keyboard coverage of material way past their standard warranty . it was eating away at margins. ) .
At took almost 4 years to switch off with two mediation iterative designs. Yes it became somewhat of a public relations disaster but that was more so due to the actual technical performance of the solution. No that the volume of the subset of complainers was "louder". Apple put tons of money and effort into trying to make butterfly work for two iterations. They really did want it to work. Cost probably played the biggest factor.
Original Boot Camp position really wasn't a technical fail, but strategically Windows 10 ARM is in no where near that same position as Windows x86 was in 2006. There is very little inertia there at all. Much of the complaining is about detaching from the older CPU system that is being switched away from ( e.g., to map historically it would be large complaints about how have to say on PPC because their other OS needed PPC. Running Linux-PPC or FreeBSD PPC or Power AIX instances .... that didn't happen).
Opening the door on BIOS was a backware boot hole that Apple got stuck with. Some folks have been "loudly" moaning incessantly about the T-series addition to the Macs from the their first addtion. T2 even more so and louder. And Apple Slicon .... brings an even deeper secure boot. Pushing Apple off of security isn't even close to the same as opening optional door to juggernaut that Apple can do little at the time against.
If folks complain "loudly" Apple isn't going to change their mind on GPL3. They just extremely likely are not. Loud complaints aren't going to get Apple to announce products 12 months in advance ( detailed long term road maps for everyone ). "Loud" isn't going to change that.
If point out that Apple is missing out on a rather substantial revenue opportunity they may shift. But how loudly point that out is not the material difference to them following through. .
Decent chance Apple lets in a couple of validated , Apple signed type 1 Hypervisors ( e.g., on the line of MacStadiums Orka Linux/KVM solution
MacStadium is offering 9to5Mac readers can get 50% off the first six months of its cloud-hosted Mac mini subscriptions with code WWDC2020.
9to5mac.com
) more so than Windows 10 on ARM as single instance boot solution. Right now that would probably drive at least as many Mac sales as the narrow corner cases where guest level virtualization doesn't work with 'pure' arm64 Windows apps demand.
Docker ( and its bindings to Linux/KVM) [and ESXi and a few other type 1 hypervisors ] is a bigger revenue/opportunity player in ARM space directly adjacent to the Mac right now that Windows 10 ARM probably is . And Apple explicitly mentioned Docker as something on the horizon as a issue needed to be dealt with. ( They could do layer on layer. MacOS Hypervisor -> Linux/KVM -> guest , but could be a thinner layering and still signed by Apple. )
P.S. It wouldn't be a "GOLDEN OPPORTUNITY" for Windows 10 on ARM to be most performant on a system that would not ship with Windows 10 on ARM on it so can boot out of the box. That would be as much highlighting a failure as a success for Microsoft. If nobody who wants to deliberately wants to build Windows can build a 'great one' then why jump on the OS variant at all? ( Presuming that Intel and AMD get better at doing x86_64 solutions over time. The bulks of follks on x86_64 will just stay there because that is where the inertia and good enough progress is. )
