Ina Fried, Axios:
Although the company has yet to say so publicly, developers and Intel officials have privately told Axios they expect such a move as soon as next year.
This seems inevitable, and tied to Bloomberg’s Marzipan reporting.
From that Bloomberg post:
Later this year, Apple plans to let developers port their iPad apps to Mac computers via a new software development kit that the company will release as early as June at its annual developer conference. Developers will still need to submit separate versions of the app to Apple’s iOS and Mac App Stores, but the new kit will mean they don’t have to write the underlying software code twice, said the people familiar with the plan.
So much to chew on here. One sense I do get is that of a grand master plan unfolding, all of this leading to a new ARM-based, shared code base vision. But a vision that is not quite in focus yet. I’ll be very interested to see if Apple shares the big picture at this summer’s WWDC.
I’m not sure about that timeline. But if true, they would have to announce at WWDC.
As an aside, I’ve been theorizing that the first ARM Mac might not be the low end machine that one would assume it would be, but rather, a high end machine, maybe even the mysterious new Mac Pro.
Proving that ARM is ready for prime time by going with the “big iron” first would make a big splash, which I imagine would be bad news for Intel.
Bloomberg is hardly a credible source when it comes to Apple. Especially that author. But I think Axios is right. But it will start at the bottom with the lowest-level MacBooks and work its way up to the Mac Pro in a few years. Apple has shown they can design better processors than Intel can. (And AMD.) Not just low-powered ones. Processors that fit into their ecosystem. I would not be surprised to see them announce such an effort at WWDC. But what we won’t see is some merging of MacOS and iOS. Considering the mess that is Apple’s own first efforts with Marzipan. I suspect people will be working on Mars first.
I doubt they’ll talk at WWDC about this unless they have a machine to show. Not necessarily announce, but remember the Intel transition…? Jobs saying something like We’re doing this… and here’s something more… the machine I’ve been using? It’s running on Intel…
I am still thinking of an arm based mac mini in the form factor of an apple tv, with early models seeded at wwdc. The first models should be ones suitable for mass developer use so devs and vendors can have months before widespread launch to recompile and test apps, in some/many cases possibly using the Marzipan stuff. If it was launched in any other way than WWDC, the attendees would be justified in asking for their money back.
There are a whole lot of difficult software conversions that will come with this transition. Certain app groups will face it the most:
– Video, graphics, music, and recording tools
– Software development tools
– Industry specific solutions
If Apple is planning this transition then they would already be working with their biggest vendors like Adobe and Microsoft.
The plural of rumor is now “inevitable.”
Completely ARM-based, this soon? I can imagine ARM chips being added to new Intel Macs, to cover specific tasks. But what do I know?
“As soon as” is very convenient phrasing, of course…
Only to the extent that they use actual embedded assembly code (or similar low-level tricks dependent on very specific CPU behavior).
Anything working with any abstraction layer or at a normal high level won’t really notice.
Some graphics stuff, though almost all of it will be using a graphics API that Apple does the implementaiton for anyway, so…?
Office? Just recompile it. It should be fine. (MS already has office on ARM for Windows RT anyway!)
(Dev tools? Won’t really care – and Apple will in any case dogfood that first, well before any release to customers.
Apple uses llvm for compiling at the metal, and it already has ARM support.)
There’s no advantage to doing that and a cost disadvantage. It’s better to simply move to the new CPU, support unported apps via emulation and make it easy to port apps so that you move through the transition as quickly as possible.
The question to me is whether ARM chips could provide acceptable performance under emulation. The reason to move to Intel was CPU power. The reason to move to ARM is… what?
Far as I’ve been told, full ownership of the hardware and power economy.
There are a whole lot of difficult software conversions that will come with this transition. Certain app groups will face it the most:
If Apple is planning this transition then they would already be working with their biggest vendors like Adobe and Microsoft.
Addendum: I missed games as a category. I play Scrabble and Catan so I’m hardly a gamer. 😉
Yeah, that would be my approach too. But then again, back in 2005 I thought Apple would be more likely to use an Intel manufactured PowerPC than change instruction sets again. 🙂
lotta supposition in that one…
I just hope they introduce a USB C or Thunderbolt 3 capable drawing monitor with the Mac Pro this year. Being able to have an app like Procreate on the Mac would be fantastic!
I’d love to be able to buy an App and find that it works on my Mac, my iPad and iPhone in appropriate way
You still have the advantage. I’m no developer or engineer, but it seems as though they’d make this transition a genuine transition if they can. Unless Marzipan (ugh, that name) is intended to somehow waltz through such a thing.
Not a good reason. That only benefits Apple.
Go listen to Jobs announcing the Intel transition and the reasoning is, really, the same reasoning that they should be able to articulate now, that they want to deliver cool products in the future and can’t see the Intel roadmap enabling that.
I can’t link to it but on youtube it’s titled Apple WWDC 2005-The Intel Switch Revealed
When developers move from the Mac because node.js, java, golang, etc are not running well then it means that fewer apps get built for the Apple ecosystem. Even building iOS apps generally means that you’ll also be building server side components in another language.
It used to be that Macs were hugely popular among developers. Devs have generally disliked the new keyboards and have hated the slow upgrade cycles. Lots have moved to better hardware running Linux. When we go through a cycle where 95% of the dev tools running on the Mac go through a bumpy transition then that drops even more.
Haven’t Apple made complaints of chop availability and advancement with Intel? IF so, that would be similar to the issue they had with PowerPC, so I can definitely see them considering another switch.
Again, take a few minutes and listen to the Jobs presentation. Yes, it’s mostly that the chip roadmap can’t take them where they want to go but he outlines, in a few minutes, a little of how they thought about that.
The problem with putting desktop class ARM chips AND Intel chips in a machine is the you increase costs, increase complexity and… for what? The 2 other times they’ve moved, they moved a machine entirely to the new chip and relied on software to do the emulation. If the ARM chips can’t do that, why move? For intel it was performance and esp performance/watt. For ARM… is it the same? Something else?
My fear is that they lose sight of the reason to do this – because it allows them to build products that they can’t see their way to build with the Intel roadmap. That was the reason they moved to Intel… because they wanted to create great products and PPC wasn’t going to enable that. NOT because of operational concerns or vague “we want to control things” reasons.
The focus on services bothers me because what’s made Apple special have been the awesome, ‘insanely great’ products.
Apple seems less and less concerned every year with people who use computers to do actual work.
If you’re running node.js and java for your app that isn’t running on a web server, you’re already in a world of suck and despair.
(Golang? Do you mean slow *compiles*, or slow *binaries* because of a bad Go compiler?)
Seems to me like you’re talking about people doing web-dev, not people doing native OSX dev, which is what everyone else is talking about?
Apple’s never going to make that product.
Wacom will sell you one right now, though.
(That and, well, why the sheer assumption that an Intel-class ARM core can’t run any of that well?
ARM’s already moving into server space, where Java and Node.js are the thing that matters, eh?
[God help us all that either of those statements are true, but that’s another rant.])
You make good points Sigivald. I luckily get to work lots of developers world wide who are all better than me (everyone breath a sigh of relief).
Talking with them and people in other tech companies I hear a few things.
Basically it adds up to headaches. Will they work it out? Sure. On the plus side, I haven’t deployed a “server” in 2-3 years because we’ve all switched to containers, etc. Docker, Kubernetes, etc is the norm now.
BTW, Docker is written in Golang so we all hope that will run well locally – it probably won’t for a while. Like other devs I’ve pushed a lot of my workload to server using tools like Terraform (Golang again) to automate it.
I sincerely hope this goes smoothly but I’ve seen transitions from Niagara to Intel, PowerPC to Intel (I’m looking at you XBox dudes), and even Android issues on various flavour on ARM. It’s never smooth.
It’s not all doom and gloom but about 90% of devs I know used to use Macs as their primary development platform. Unscientifically it now feels like 70-75%. If the number goes down again it has an impact on the overall ecosystem.
And I likes my Apple ecosystem so I hope it’s a smooth transition.
True, developer devices would have to be made available before the official launch, but I wouldn’t call such devices as official models, the same way there were developer boxes available during the year between the announcement of, and the shipping of, the new intel macs over a decade ago.