What Apple Announced at WWDC 2020
It looks like I got some things right, and some things wrong, in my previous post. Let’s look at what Apple actually announced.
Mac with Apple Processors, across the line
Yes, they really are transitioning the whole Mac line to Apple’s own processors. The timeline they announced is “within 2 years” to transition the whole line. I still expect that they’ll end up well within two years, maybe even closer to a year. Similar to how the Intel transition was announced as “a year and a half”, but ended up being shorter.
The first ARM Macs will be available this year. This was a surprise to a lot of pundits, but it makes total sense to me, given that several major third-party software vendors (Adobe, Microsoft, and Epic) are on-board with the switch, and have their software working already. I was expecting both Adobe and Microsoft to show working pre-release software, just because it really is that easy to move a modern Mac code base to ARM, and they’ve both recently gone through a fairly-painful 64-bit transition for Catalina.
Rosetta 2, to run existing Intel Mac Applications
The big surprise for me is that they did include x64 compatibility in Big Sur. I’m happy to be wrong about that, it’s obviously good news for users. I just figured that the chance to make a clean break would be very tempting to Apple.
Rosetta 2 uses a combination of translation at install time for applications, and translation at load time for plugins. I think that ahead of time translation is a good tradeoff between taking some extra time to get started, in exchange for getting better translation. JIT translation of machine code is hard to balance between performance and latency.
The Rosetta 2 documentation is pretty sparse right now, but I did get the impression that x64-JIT compilers are supported in Rosetta apps, which is interesting. Presumably, when you make the syscall to make a segment of code executable, they translate it then. Pretty slick, though I wonder how much it’ll cause performance hiccups in, for example, web browsers, which rely heavily on JIT to get adequate performance.
Running iPad and iPhone applications without modification
Another thing that seems to have taken a bunch of people by surprise is that ARM Macs will be able to run iPad and iPhone apps without any modifications. This is a logical outgrowth of the Catalyst software that lets you rebuild iPad apps to run on the Intel version of MacOS. You just don’t need to recompile on the ARM Mac, because they’re already the same processor architecture.
Just like existing Catalyst applications, it’s possible for developers to add Mac-specific features (e.g. menus), to create a better user experience on the Mac. This really does make UIKit (or SwiftUI) the “obvious” choice for new application development.
Porting assistance for Open-Source apps and frameworks
An interesting news item that came out of WWDC is that Apple is going to make patches available to a number of Open Source projects, to help get them up and running on ARM Macs. This includes Chromium, Node.js, OpenJDK, and Electron, which should mean that those projects won’t lag far behind.
So, what’s it all mean?
To me, this seems like just about entirely a win. The new Macs will be faster, use less power, will have a larger software library (all those iOS apps), and more capabilities.
Some software will get “left behind” in the transition, but not very much, at least in the short term. Running software under Rosetta will likely not be great, performance-wise, but it’ll be adequate for a lot of uses.
There is one major downside, for a certain subset of users
No x64 processor means no Bootcamp support for booting into WIndows, and no virtualization software to run your other x86 operating systems. I have a friend who uses Docker extensively as part of his developer workflow, and he’s just going to be out of luck, as far as using a Mac goes.
There is virtualization support on the ARM Macs, but it’s for virtualizing other ARM operating systems. You’ll be able to run an ARM Linux OS inside Parallels (for example), but if your workflow right now includes running code in x64 Windows or Linux, the ARM Macs won’t be a good fit.
What about Games?
It seems like every time a major new MacOS version comes out, they claim it’s going to be “great for games”, but the games mostly don’t actually come.
Having Unity supporting ARM Macs will definitely make it easier for anyone already using Unity to support the new Macs. But the current version of Unity already supports Mac, and still a lot of games never make it there, so I don’t think that’s a win. If anything, it’s a loss, since anybody who wants to support the Mac at least needs to test with both Intel and ARM Macs.
While a lot of big-name games never make it to the Mac, there’s actually quite robust support for the Mac among small indie games developers on Steam and Itch. Again, some of these folks will look at the cost to support both kinds of Macs, and decide it’s not worth it, so we’ll probably lose a few there, as well.
But then there are iPad games. A lot of iPad games are “casual games”, which is just fine by me, since I play more of that sort of thing than I do first-person shooters. And given that iPad games will, by and large “just work” on the new Macs, we may see more iPad games developers moving a bit more upscale. It’ll be interesting to see.
WIll I buy one?
We’ll see what gets announced, but I expect that I will, whenever the “pro” laptops are available.